1 INFO-VAX	Fri, 19 Apr 2002	Volume 2002 : Issue 216       Contents: Re: Alpha tuning didn't  ANN: HGFTP V3.1  Re: ANN: HGFTP V3.1 , Re: Anyone using CHARON-VAX for development?, Re: Anyone using CHARON-VAX for development?1 Re: Availability Manager not displaying all nodes 1 Re: Availability Manager not displaying all nodes 1 Re: Availability Manager not displaying all nodes 1 Re: Availability Manager not displaying all nodes 1 Re: Availability Manager not displaying all nodes ) Re: Best DCL programm you've ever written ) Re: Best DCL programm you've ever written ) Re: Best DCL programm you've ever written ) Re: Best DCL programm you've ever written  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: Buffered IO performance  Re: Buffered IO performance % Console scan profile for Gigaswitches   Re: CVS V1.11.2 on OpenVMS Error  Re: DEC 3000/400 VMS 6.2 problem+ Emulate "show calls" in application logging / Re: Emulate "show calls" in application logging   Re: Fearless VMS Prognostication  Re: Fearless VMS Prognostication  Re: Fearless VMS Prognostication  Re: Fearless VMS Prognostication  Re: Fearless VMS Prognostication Re: File IO query. FLAME ALERT Re: File IO query. FLAME ALERT- Re: getting SIMH 2.9-6/VAX + VMS configured ? - Re: getting SIMH 2.9-6/VAX + VMS configured ? ) Re: How to prevent smtp relay w/Multinet? ) Re: How to prevent smtp relay w/Multinet? ) Re: How to prevent smtp relay w/Multinet? ) Installing Netscape Gold from Hobbyist CD  Re: Itanium troubles> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)0 Re: linking with pthread library for C++ program0 Re: linking with pthread library for C++ program4 Lleg? la oportunidad que esperabas... CAMBI? TU VIDA- Mails with long lines (> 255 bytes) truncated 1 Re: Mails with long lines (> 255 bytes) truncated 1 Re: Mails with long lines (> 255 bytes) truncated 1 Re: Mails with long lines (> 255 bytes) truncated 1 Re: Mails with long lines (> 255 bytes) truncated 1 Re: Mails with long lines (> 255 bytes) truncated B Re: Major Virus Alert!  Better get on VMS or APPLE now or bye bye!3 Re: MicroVAX 3100-30 self test failure: RT-Retrbctr  Mime hole virus ) OpenVMS E7.3-1 changes show dev d display - Re: OpenVMS E7.3-1 changes show dev d display - Re: OpenVMS E7.3-1 changes show dev d display < Patch *entitlement*: what does *entitlement* actually mean ?) Re: Predictions - just for the hell of it ) Re: Predictions - just for the hell of it  Simh V2.9 problem  Re: SSH for Alpha OVMS? 8 Re: Stuck in licenseless Motif with OpenVMS 7.2 Hobbyist8 Re: Stuck in licenseless Motif with OpenVMS 7.2 Hobbyist8 Re: Stuck in licenseless Motif with OpenVMS 7.2 Hobbyist8 Re: Stuck in licenseless Motif with OpenVMS 7.2 Hobbyist0 Re: Sue, a great promotion for "unhackable" vms! Re: System Disk as Quorum Disk" TCP/IP 5.0 Configuration Questions, Re: Time to finally rename VMS back to VMS ?, Re: Time to finally rename VMS back to VMS ?, Re: Time to finally rename VMS back to VMS ? Two DMQ COM servers running  Re: Two DMQ COM servers running  Re: Two DMQ COM servers running , Re: UNXSIGNAL bugcheck on virtual disk, why?, Re: UNXSIGNAL bugcheck on virtual disk, why? Re: VAX/Alpha CI Re: VAX/Alpha CI Re: VAX/Alpha CI Re: VAX/Alpha CI' Re: VMS Cluster address switching issue ' Re: VMS Cluster address switching issue  [Fwd: New OpenVMS Whitepaper]   F ----------------------------------------------------------------------  % Date: Fri, 19 Apr 2002 11:31:01 -0400 ; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com>   Subject: Re: Alpha tuning didn't$ Message-ID: <3cc03896$1@news.si.com>  A >> set RECNXINTERVAL 180                  |  set RECNXINTERVAL 20  > # >Are you experiencing LAN problems?   K On occasion, some device on the net doesn't play nice and sends a storm.  I " don't want to lose the satellites.  < >> set LOCKDIRWT 0                        |  set LOCKDIRWT 0 > D >So all lock requests for shared files will be remote.  Is that what >you intended?   Yes.  B >> set SHADOW_MAX_COPY 0                  |  set SHADOW_MAX_COPY 1 > ( >Do you use host-based volume shadowing?   No shadowing of any kind.  --A Brian Tillman                   Internet: tillman_brian at si.com A Smiths Aerospace                          tillman at swdev.si.com = 3290 Patterson Ave. SE, MS      Addresses modified to prevent < Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  # Date: Fri, 19 Apr 2002 13:54:16 GMT - From: goathunter@goatley.com (Hunter Goatley)  Subject: ANN: HGFTP V3.10 Message-ID: <3cc00ce1.77621944@news.process.com>  ) HGFTP V3.1 is now available for download.   A I hadn't intended another release so soon, but I tracked down the = problem between HGFTP and the Opera web browser and wanted to  release the fix.  8 This release includes the following changes from V3.0-1:   New:C    - The HGFTP Listener now provides the ability to restrict access D      to the FTP server based on the IP address of the remote client.F      Two new logicals, HG_FTP_ALLOW_IP_LIST and HG_FTP_REJECT_IP_LIST,F      can be defined to allow or reject connections based on IP address      or network address.   Fixed:A    - HGFTP works with the Opera web browser now.  Opera was using >      MODE B (block mode) for file transfers, and block mode isC      apparently broken in HGFTP.  (Block mode is rarely used---this B      is the first time since the release of HGFTP 8 years ago that>      I've heard of a client actually trying to use it.)  Block@      mode is disabled in the HGFTP server in V3.1, and Opera now!      works with the HGFTP server.   D    - The HGFTP client has always converted any initial command givenB      on the DCL command to lowercase.  This was also affecting the@      file specification in URLs (with the /URL qualifier)---URLsD      with case-sensitive file specs could not be used.  This problemC      has been corrected, for both /URL handling and for the initial '      command given on the command line.   8 HGFTP can be downloaded using any of the following URLs:   http://www.process.com/openvms/   5 ftp://ftp.process.com/vms-freeware/fileserv/hgftp.zip : http://vms.process.com/ftp/vms-freeware/fileserv/hgftp.zip1 ftp://ftp.tmk.com/vms-freeware/fileserv/hgftp.zip 6 http://www.tmk.com/ftp/vms-freeware/fileserv/hgftp.zip  2 And on the other mirrors within the next 24 hours.   Hunter ------9 Hunter Goatley, Process Software, http://www.process.com/ 8 goathunter@goatley.com    http://www.goatley.com/hunter/< New Robert R. McCammon site: http://www.RobertRMcCammon.com/   ------------------------------  # Date: Fri, 19 Apr 2002 16:16:54 GMT - From: goathunter@goatley.com (Hunter Goatley)  Subject: Re: ANN: HGFTP V3.10 Message-ID: <3cc04213.91239645@news.process.com>  P On Fri, 19 Apr 2002 13:54:16 GMT, goathunter@goatley.com (Hunter Goatley) wrote:  * >HGFTP V3.1 is now available for download. > ; Sigh.  A last minute documentation correction resulted in a 9 bogus documentation saveset in the kit.  I've updated the ; .ZIP file in place.  If you downloaded before 12:10 PM EDT, 7 then please download the file again.  (Or don't install ; the documentation; if you answer no to installing the docs,  HGFTP will install OK.   My apologies for the oversight.   8 HGFTP can be downloaded using any of the following URLs:   http://www.process.com/openvms/   5 ftp://ftp.process.com/vms-freeware/fileserv/hgftp.zip : http://vms.process.com/ftp/vms-freeware/fileserv/hgftp.zip1 ftp://ftp.tmk.com/vms-freeware/fileserv/hgftp.zip 6 http://www.tmk.com/ftp/vms-freeware/fileserv/hgftp.zip  2 And on the other mirrors within the next 24 hours.   Hunter ------9 Hunter Goatley, Process Software, http://www.process.com/ 8 goathunter@goatley.com    http://www.goatley.com/hunter/< New Robert R. McCammon site: http://www.RobertRMcCammon.com/   ------------------------------    Date: 19 Apr 2002 06:40:13 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) 5 Subject: Re: Anyone using CHARON-VAX for development? 3 Message-ID: <DZ105NJnrujT@eisner.encompasserve.org>   \ In article <3CBF37F6.11794.142A17E9@localhost>, "Stanley F. Quayle" <stan@stanq.com> writes:. > On 18 Apr 2002, at 9:22, Keith Parris wrote:F >> Why couldn't the emulator detect the special case of an instruction  >> branching to its own address? > D > Because that could happen anywhere.  Imagine a piece of user mode F > code that is executing a spinloop.  You don't want to slow down the 4 > emulator at the user level, only the kernel level.  E If the loop is counting anything, it does not match the criteria that D was proposed (jump to self).  The VMS null process is different from a spinlock in this regard.  F If the loop is waiting for a timer interrupt, there is no need for theD emulator to stop its clocking operations, just the emulation of jump to self.   ------------------------------    Date: 19 Apr 2002 10:51:21 -07001 From: KeithParris_NOSPAM@yahoo.com (Keith Parris) 5 Subject: Re: Anyone using CHARON-VAX for development? = Message-ID: <6ec1251e.0204190951.2f3f283a@posting.google.com>   a "Stanley F. Quayle" <stan@stanq.com> wrote in message news:<3CBF37F6.11794.142A17E9@localhost>... . > On 18 Apr 2002, at 9:22, Keith Parris wrote:G > > Why couldn't the emulator detect the special case of an instruction ! > > branching to its own address?  > D > Because that could happen anywhere.  Imagine a piece of user mode F > code that is executing a spinloop.  You don't want to slow down the 4 > emulator at the user level, only the kernel level.  ? It doesn't really matter where a branch-to-self occurs, even in D user-mode code.  All the system would do is allow the process to eat@ up CPU until the next quantum end anway.  Might as well have theC emulator let other Windows applications do work until that happens. . ----------------------------------------------. Keith Parris | parris at encompasserve dot org   ------------------------------  % Date: Fri, 19 Apr 2002 08:33:31 +0200 ) From: Bart Zorn <B.Zorn@xs4all.nospam.nl> : Subject: Re: Availability Manager not displaying all nodes/ Message-ID: <3CBFBA3B.1090000@xs4all.nospam.nl>    Shawn Joseph wrote: G > I have installed Availibility Manager on 3 nodes.  I can only see one # > of them in the Windows Java Tool.  > G > The one I can see is Alpha VMS version 7.2-1H1 and is by itself.  The F > other 2 nodes are in a cluster and are Alpha VMS 7.2-1.  I installed5 > Availability Manager 2.0 on all three nodes and ran H > @SYS$STARTUP:AMDS$STARTUP START on all three nodes and it said startupF > was successful on all three nodes.  However only the standalone nodeE > shows up in my Windows tool.  I remember seeing something about the G > Java runtime environment needing to be a certain version, but I think H > it also said that AM came with it's on JVM that was installed with theF > product.  How would I determine what version of Java is installed on > each of these machines?  > G > It also says to run avail/avail to start the data analyzer. Does this A > need to be done from DecWindows or something?  I keep getting a B > message about not being able to connect to an X11 window server. > A > I tried to find a process on the system for AM after successful G > startup and couldn't find anything relating to this, does anyone know % > the name of the process it creates?  > E > I was also looking for some kind of log files to help me figure out 6 > why I can't see the other nodes but didn't find any.  F Are all the relevant systems on the same network broadcast domain? In E other words, is there a router or a switch with filtering in between  E them? Availability manager uses it's own network protocol, different   from IP or DECnet or whatever.  	 Bart Zorn    ------------------------------  % Date: Fri, 19 Apr 2002 08:45:06 -0500 1 From: "Dave Gudewicz" <david.gudewicz@abbott.com> : Subject: Re: Availability Manager not displaying all nodes8 Message-ID: <a9p734$5fq$1@fizban.fizban.pprd.abbott.com>  G Are all the node on the same LAN?  If not, then problems like this will  occur.  0 Same wire only for AM.  Fixed in future release.   Dave...   4 "Shawn Joseph" <sdk_joseph@msn.com> wrote in message6 news:f897700f.0204181335.544b5c6@posting.google.com...G > I have installed Availibility Manager on 3 nodes.  I can only see one # > of them in the Windows Java Tool.  > G > The one I can see is Alpha VMS version 7.2-1H1 and is by itself.  The F > other 2 nodes are in a cluster and are Alpha VMS 7.2-1.  I installed5 > Availability Manager 2.0 on all three nodes and ran H > @SYS$STARTUP:AMDS$STARTUP START on all three nodes and it said startupF > was successful on all three nodes.  However only the standalone nodeE > shows up in my Windows tool.  I remember seeing something about the G > Java runtime environment needing to be a certain version, but I think H > it also said that AM came with it's on JVM that was installed with theF > product.  How would I determine what version of Java is installed on > each of these machines?  > G > It also says to run avail/avail to start the data analyzer. Does this A > need to be done from DecWindows or something?  I keep getting a B > message about not being able to connect to an X11 window server. > A > I tried to find a process on the system for AM after successful G > startup and couldn't find anything relating to this, does anyone know % > the name of the process it creates?  > E > I was also looking for some kind of log files to help me figure out 6 > why I can't see the other nodes but didn't find any. >   > Any help would be appreciated. >  > Respectfully,  > Shawn Joseph   ------------------------------  % Date: Fri, 19 Apr 2002 10:02:45 -0400 1 From: "Mark D. Jilson" <jilly@clarityconnect.com> : Subject: Re: Availability Manager not displaying all nodes2 Message-ID: <3CC02385.DE450B83@clarityconnect.com>  H If the nodes that do not show up have more than one nic or other networkG interface then you need to edit AMDS$LOGICALS.COM and point AMDS$DEVICE D to the nic that will allow the AMDS packets to be transmitted to theC PC.  Additionally make sure that this network path is configured to G allow the AMDS packets to travel to the PC.  The AMDS docs have all the # info your network admins will need.    Shawn Joseph wrote:  > G > I have installed Availibility Manager on 3 nodes.  I can only see one # > of them in the Windows Java Tool.  > G > The one I can see is Alpha VMS version 7.2-1H1 and is by itself.  The F > other 2 nodes are in a cluster and are Alpha VMS 7.2-1.  I installed5 > Availability Manager 2.0 on all three nodes and ran H > @SYS$STARTUP:AMDS$STARTUP START on all three nodes and it said startupF > was successful on all three nodes.  However only the standalone nodeE > shows up in my Windows tool.  I remember seeing something about the G > Java runtime environment needing to be a certain version, but I think H > it also said that AM came with it's on JVM that was installed with theF > product.  How would I determine what version of Java is installed on > each of these machines?  > G > It also says to run avail/avail to start the data analyzer. Does this A > need to be done from DecWindows or something?  I keep getting a B > message about not being able to connect to an X11 window server. > A > I tried to find a process on the system for AM after successful G > startup and couldn't find anything relating to this, does anyone know % > the name of the process it creates?  > E > I was also looking for some kind of log files to help me figure out 6 > why I can't see the other nodes but didn't find any. >   > Any help would be appreciated. >  > Respectfully,  > Shawn Joseph   --  C Jilly	- Working from Home in the Chemung River Valley - Waverly, NY 0 	- jilly@clarityconnect.com			- Brett Bodine fan. 	- Mark.Jilson@Compaq.com			- since 1975 or so, 	- http://www.jilly.baka.com               -   ------------------------------    Date: 19 Apr 2002 07:08:20 -0700' From: sdk_joseph@msn.com (Shawn Joseph)V: Subject: Re: Availability Manager not displaying all nodes= Message-ID: <f897700f.0204190608.3938319c@posting.google.com>   c rob.buxton@wcc.govt.nz (Rob Buxton) wrote in message news:<3cbf63ae.2491225613@news.wcc.govt.nz>... B > On 18 Apr 2002 14:35:57 -0700, sdk_joseph@msn.com (Shawn Joseph) > wrote: >   > Could be a couple of problems. > G > From  your Post I'm guessing that you're running the AM Frontend from 	 > a P.C.?D(            Yes I am running it from a PC > 4 > In the left hand screen, you have a group DECAMDS?E > And if you highlight this you only see one system in the right hand  > panel?            Yes that is correct > G > It may be that the version of the RMDriver is not compatible with thehC > version supplied by AMDS. If AMDS has replaced the version of thee6 > RMDriver it may require a reboot to get this loaded.? > I would have thought that there was a message indicating thisS > somewhere though. @              I have rebooted the system(s) about 6 times since I installed the software.  > C > To separate systems, you may want to change the definition of thee2 > AMDS$GROUP_NAME Logical for the Cluster Members. > B > The only other thing I can think of is that DECAMDS doesn't have > access to the nodes.< > On the different VMS Systems there needs to be a line like > *\1DECAMDS\RE > in the SYS$COMMON:[SYSMGR]AMDS$DRIVER_ACCESS.DAT file, in this caseeA > 1DECAMDS needs to match the password value in the Customize VMSe& > section, under Customize on the GUI.  D    I have checked this per your recommendation and verfied the front! end and back end passwords match.  > H > >I have installed Availibility Manager on 3 nodes.  I can only see one$ > >of them in the Windows Java Tool. > >bH > >The one I can see is Alpha VMS version 7.2-1H1 and is by itself.  TheG > >other 2 nodes are in a cluster and are Alpha VMS 7.2-1.  I installedb6 > >Availability Manager 2.0 on all three nodes and ranI > >@SYS$STARTUP:AMDS$STARTUP START on all three nodes and it said startuplG > >was successful on all three nodes.  However only the standalone node2F > >shows up in my Windows tool.  I remember seeing something about theH > >Java runtime environment needing to be a certain version, but I thinkI > >it also said that AM came with it's on JVM that was installed with therG > >product.  How would I determine what version of Java is installed ont > >each of these machines? > >3H > >It also says to run avail/avail to start the data analyzer. Does thisB > >need to be done from DecWindows or something?  I keep getting aC > >message about not being able to connect to an X11 window server.t > >tB > >I tried to find a process on the system for AM after successfulH > >startup and couldn't find anything relating to this, does anyone know& > >the name of the process it creates? > >sF > >I was also looking for some kind of log files to help me figure out7 > >why I can't see the other nodes but didn't find any.p > > ! > >Any help would be appreciated.o > >f > >Respectfully, > >Shawn Josephs  B I also ensured all nodes and the front end workstation were on the2 same network segment with the same broadcast mask.   ------------------------------  % Date: Fri, 19 Apr 2002 16:49:28 +0200M) From: Bart Zorn <B.Zorn@xs4all.nospam.nl>v: Subject: Re: Availability Manager not displaying all nodes- Message-ID: <3CC02E78.10608@xs4all.nospam.nl>    Shawn Joseph wrote:ke > rob.buxton@wcc.govt.nz (Rob Buxton) wrote in message news:<3cbf63ae.2491225613@news.wcc.govt.nz>...  >  	[Snip]u > D > I also ensured all nodes and the front end workstation were on the4 > same network segment with the same broadcast mask.  I The first part of this sentence is OK, but not the second. AM (and AMDS) nH use their own LAN protocol which has nothing to do with IP or broadcast  masks.  	 Bart Zorn    ------------------------------  % Date: Fri, 19 Apr 2002 12:16:29 +0100t* From: "Rainer Giese" <giese@volkswerft.de>2 Subject: Re: Best DCL programm you've ever written5 Message-ID: <a9oqpu$51jas$1@ID-138444.news.dfncis.de>t  D Is this the place to write down what one has made the last time ?:-)  I I've made a search engine for our intranet with DCL. The spider uses LYNX I for grabbing pages and links. The user interface allows word expressions. G The daily update job take about 5 min on an ES40 for appr. 350 pages. AnI search request takes less than a second for single words and 2..4 secondsn for expressions.  L The whole matter consists of 10 procedures with 1300 line and a 75000 record word index file.F Btw., I'm able to grab PDF-Files with an converter, but anyone knows a! MS-Word-to-TXT-Converter on VMS ?    Regards,   Rainer Giese   ------------------------------  % Date: Fri, 19 Apr 2002 04:48:01 -0600,' From: Thomas Dzubin <dzubint@vcn.bc.ca>t2 Subject: Re: Best DCL programm you've ever written% Message-ID: <3CBFF5E1.1CAE@vcn.bc.ca>n   Gijs wrote:>B > When was it, what does it do for you and whe can it be obtained?  C I wrote an integer BASIC interpreter about ten years ago (1990-ish) A entirely in DCL...it was pretty simple, but it worked and I stillr occasionally use it for fun.  A I know nearly everybody here has written a "change directory" DCLeG command, and I'm no exception...I'm still using the one that I wrote ine+ 1988 (it knows various Unix-isms like "..")lH I also wrote a "finger" command since I didn't like the output format of< either SHOW USERS or SHOW SYSTEM ...I still use finger also.  H Sadly, I don't do much DCL scripting anymore...that part of my brain hasH been migrated to Perl and most of my quick&dirty stuff is now written in Perl instead of DCL.  
 Thomas Dzubin ' Vancouver, Calgary, or Saskatoon CANADAi   ------------------------------  % Date: Fri, 19 Apr 2002 12:28:20 +0100r( From: Nic Clews <sendspamhere@127.0.0.1>2 Subject: Re: Best DCL programm you've ever written) Message-ID: <3CBFFF54.DF1DB168@127.0.0.1>G   Gijs wrote:  > B > When was it, what does it do for you and whe can it be obtained?  ? I guess I have to put an entry in for my PAPER.COM, a 200 block  monstrosity that:o  G 1. Used accounting to determine paper usage over a seven day period perp user account3 2. Send warning emails to those approaching limits.bH 3. Disable the PRINT command for those who have exceeded the 7 day paper quota / (and we made them pay a fine to be re-enabled!)iC 4. Produce a top ten list of the worst paper abusers for display at/ system login6 (The hall of shame but didn't always have that effect)= 5. Automatically re-enable those after their exclusion period/  7 No, you don't want a copy, and it was in the late 80's.f  E Anyone reading this list become a victim of this? If you've read this-+ you'll know where it was! And yes, it's me!/   -- o( Regards, Nic Clews CSC Computer Sciences nclews at csc dot come   ------------------------------  # Date: Fri, 19 Apr 2002 12:14:22 GMT-1 From: LESLIE@JRLVAX.HOUSTON.RR.COM (Jerry Leslie)12 Subject: Re: Best DCL programm you've ever written: Message-ID: <ySTv8.17971$o86.704843@typhoon.austin.rr.com>  ) Rainer Giese (giese@volkswerft.de) wrote:rH : Btw., I'm able to grab PDF-Files with an converter, but anyone knows a# : MS-Word-to-TXT-Converter on VMS ?. :   =   http://vms.process.com/scripts/fileserv/fileserv.com?CATDOCo   Hunter Goatley's FILESERVt  
   catdoc.zip aA   Description:  Extracts contents of MS-Word files as plain text T   Version:  V3.4, 4-JUN-1998 l)   Author:  Victor Wagner, Hunter Goatley t   Architecture:  VAX,AXP s   Size:  42 blocks G   Language:  C O  H --Jerry Leslie   leslie@clio.rice.edu  (my opinions are strictly my own)9   Note: leslie@jrlvax.houston.rr.com is invalid for email    ------------------------------   Date: 19 Apr 2002 08:09:01 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren)  Subject: Re: Blade architectures0 Message-ID: <a9ojat$ee5$1@pegasus.csx.cam.ac.uk>  - In article <a9o0m8$cfh@web.eng.baileynm.com>,r' Peter da Silva <peter@abbnm.com> wrote:t1 >In article <a9fa8d$gpe$1@pegasus.csx.cam.ac.uk>,/* >Nick Maclaren <nmm1@cus.cam.ac.uk> wrote:H >> If you are doing that, WHAT operating system are you running?  I want* >> it now!  Yesterday.  Pretty please .... >-& >I said "almost always", not "always". > J >> Frankly, half of the hardware crashes that I see ARE software problems. >NH >That depends on whether "Driver that doesn't properly handle revision AA >of this ethernet card between ECO 23 and 25" or "Missing obscureuG >initialization sequence in motherboards having serial numbers startingrB >with 34Q2" or "RAM that produces parity errors under UNIX but notE >Windows because it was only ever tested with Windows access patternspI >and you could probably fix it by changing the page allocation algorithm"i4 >count as "software problems" or "hardware crashes". >hE >> If the software wasn't such a pig's ear, it would have handled thegG >> (often isolated and well-defined) hardware problem and allowed me to F >> at least run the diagnostic tools that the system comes with (well,' >> it does come with them, doesn't it?)h > N >Seems like you'd call them "software problems". Fair enough. I don't considerL >these worth tracking down, given the quality of PC hardware, because you'll >never run into them twice._  @ None of the above.  Remember that I am not referring to personal
 computers.  A It is things like recovery from transient errors (memory, router,m> device, disk etc.)  The systems I manage have a LOT of code to? recover (and claim it, vociferously), and it generally works in2@ the simple cases.  When you get particularly nasty combinations,@ however, it is quite common for the logic to fail and the system to die.e  A Now, because such code is usually running VERY highly privileged,o@ and is often in firmware, the consequences of failure tend to be= dire.  Sometimes bad enough to cause things like a refusal toeA respond to the power off button!  Yes, I HAVE had to switch off ad& computer drawing 30 KW at the wall :-(  A Now, I know that you were referring to personal computers, but my @ assumption is that the same almost certainly holds true on them.A Yes, the root case may be hardware, and the failure may look like-> hardware, but the fact that the failure is so horrible is veryC likely to be software.  After all, an increasing number of personalaD computers are running the same software as is used on large systems.     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679    ------------------------------  % Date: Fri, 19 Apr 2002 10:29:34 +0100 4 From: Mark Horsburgh <M.K.Horsburgh@damtp.cam.ac.uk>  Subject: Re: Blade architecturesD Message-ID: <slrnabvoru.igd.M.K.Horsburgh@krull.dialup.ntlworld.com>  J On Fri, 19 Apr 2002 15:26:08 +1200, John Homes <john.homes@eds.com> wrote: > 5 > "Peter da Silva" <peter@abbnm.com> wrote in messagen) > news:a9o05n$bru@web.eng.baileynm.com...u: >> In article <slrnabpt5o.e4q.shannon@news.widomaker.com>,A >> Charles Shannon Hendrix  <shannon@nospam.widomaker.com> wrote:tL >> > How many systems out there try or have tried to do virtual memory whereI >> > the executable file is part of mix, in a system where a code page is M >> > just marked free and later refetched from the exec instead of from swap?i >>8 >> I'm surprised that any of them *don't* work that way. >> > N > It's impossible to do it if the executable needs massaging in some way to be	 > loaded.a  H Not impossible at all. OS/2 did this, I believe. It's just that you needK to do the massaging when you fault the page back in. Presumably you have toiH do this when it's first faulted in anyway (assuming pure demand paging),J so you have to have that capability anyway. In fact, I've never understoodG why modern Unices pay the runtime cost for position independent code intC order to avoid the need to do relocation at load (page fault) time.o  	 <snipped>o  N > There is a lot of other stuff within an MVS executable that the loader needsM > to play with as part of the load. So the virtual memory subsystem cannot doz > it.d  F There is no reason in principle that the VM subsystem couldn't talk to the loader.s   Mark   ------------------------------  # Date: Fri, 19 Apr 2002 14:20:43 GMTa+ From: Anne & Lynn Wheeler <lynn@garlic.com>   Subject: Re: Blade architectures) Message-ID: <u8z7ju5xf.fsf@earthlink.net>S  ) "John Homes" <john.homes@eds.com> writes:sN > It's impossible to do it if the executable needs massaging in some way to be	 > loaded.t >aM > E.g. relocations on S/360 and its sucessors. As I recall, VM enabled you to M > build executables (called something else, but I recall not what) that couldtK > be used in that way but the price was that the (virtual) load address wasi > fixed at build time. >eN > There is a lot of other stuff within an MVS executable that the loader needsM > to play with as part of the load. So the virtual memory subsystem cannot doo > it.h  E RLD entries ... relocatable load directory ... basically locations ina: the program that have addresses that are related to an ESDD ... external symbol directory. The loader resolves the ESD addresses$ then goes after all the RLD entries.  F MVS didn't originate as a virtual memory system ... it originated as aD real storage system ... that eventually got everything laid out in aC virtual address space ... even to the paradigm that the kernel codenF and the application code co-exists in the same address space (althoughC that has slowly mitigated over the years). For years, MVS struggledu> with various kernel &/or services code getting larger & largerA ... impinging on the remaining address space left to applicationsl% (especially with 24bit/16mbyte area).o  E CMS was a single user interactive monitor that ran in virtual addressnB spaces provided by CP/67 ... but used a lot of OS/360 applicationsF (most of the compilers, applications, etc). For both CP/67 and VM/370,B CMS had certain high use code that was "shared" ... i.e. same pageF occured in multiple different address spaces (aka reduces real storageE & paging requirements) ... however these originally were single fixed  definition.t  D I wrote some enhancements to CMS & VM/370 that supported effectively? shared libraries ... additional shared code/pages that could beDB introduced in virtual address space "on the fly". I even sanitizedA some amount of CMS & CMS application code to live in shared pagesv? ... and did it in such a way that it eliminated all RLD entriesi< (i.e. the same shared code could live at different addressesE simultaneously in different virtual address spaces). The CMS code wase@ eventually shipped as original designed/written. However, only a@ subset of the VM/370 kernel support eventually shipped ... whichA didn't include the ability to position shared pages at arbritraryo& locations in different address spaces.  D The scenario is that in 24bit/16megabyte address space ... it wasn't= that frequent that any single process needed a combination of C (different) shared library code that exceeded the process's virtual E address space. However, it was the case with global fixed address per C shared library ... that an installation would find that it couldn'thC choose a unique virtual address range for every shared library thatlF their user community might want. The result was that installations hadA to eventually assign duplicate addresses to some shared librariescD ... meaning that there were certain combinations of applications (at6 least in shared library format) that weren't possible.  . past postings giving RLD & ESD format details:W http://www.garlic.com/~lynn/2001.html#8 finding object decks with multiple entry pointseW http://www.garlic.com/~lynn/2001.html#14 IBM Model Numbers (was: First video terminal?) d http://www.garlic.com/~lynn/2001.html#60 Text (was: Review of Steve McConnell's AFTER THE GOLD RUSH)5 http://www.garlic.com/~lynn/2001c.html#87 "Bootstrap"iV http://www.garlic.com/~lynn/2001k.html#31 Is anybody out there still writting BAL 370.S http://www.garlic.com/~lynn/2001m.html#45 Commenting style (was: Call for folklore):  " past floating shared segment refs:^ http://www.garlic.com/~lynn/93.html#8 PowerPC Architecture (was: Re: PowerPC priced very low!)] http://www.garlic.com/~lynn/95.html#9 Cache and Memory Bandwidth (was Re: A Series Compilers)0E http://www.garlic.com/~lynn/96.html#4a John Hartmann's Birthday Partyo@ http://www.garlic.com/~lynn/2000.html#18 Computer of the centuryD http://www.garlic.com/~lynn/2000.html#75 Mainframe operating systems9 http://www.garlic.com/~lynn/2000b.html#29 20th March 2000 W http://www.garlic.com/~lynn/2000b.html#50 VM (not VMS or Virtual Machine, the IBM sort) G http://www.garlic.com/~lynn/2000b.html#54 Multics dual-page-size schemehG http://www.garlic.com/~lynn/2000b.html#55 Multics dual-page-size scheme ? http://www.garlic.com/~lynn/2000c.html#76 Is a VAX a mainframe? S http://www.garlic.com/~lynn/2001b.html#83 Z/90, S/390, 370/ESA (slightly off topic)uR http://www.garlic.com/~lynn/2001c.html#2 Z/90, S/390, 370/ESA (slightly off topic)9 http://www.garlic.com/~lynn/2001c.html#13 LINUS for S/390hI http://www.garlic.com/~lynn/2001c.html#84 database (or b-tree) page sizesv4 http://www.garlic.com/~lynn/2001f.html#8 Theo Alkema4 http://www.garlic.com/~lynn/2001f.html#9 Theo Alkemad http://www.garlic.com/~lynn/2001i.html#20 Very CISC Instuctions (Was: why the machine word size ...)U http://www.garlic.com/~lynn/2002d.html#31 2 questions: diag 68 and calling convention ? http://www.garlic.com/~lynn/2002e.html#5 What goes into a 3090?D  " general virtual memory references:1 http://www.garlic.com/~lynn/subtopic.html#wsclockM   -- sH Anne & Lynn Wheeler   | lynn@garlic.com -  http://www.garlic.com/~lynn/    ------------------------------   Date: 19 Apr 2002 14:29:34 GMT  From: hack@watson.ibm.com (hack)  Subject: Re: Blade architectures+ Message-ID: <a9p9ke$p52$1@news.btv.ibm.com>.  , In article <a9o2oh$8be$1@hermes.nz.eds.com>,& John Homes <john.homes@eds.com> wrote: >l4 >"Peter da Silva" <peter@abbnm.com> wrote in message( >news:a9o05n$bru@web.eng.baileynm.com...: >> In article <slrnabpt5o.e4q.shannon@news.widomaker.com>,A >> Charles Shannon Hendrix  <shannon@nospam.widomaker.com> wrote:eL >> > How many systems out there try or have tried to do virtual memory whereI >> > the executable file is part of mix, in a system where a code page isgM >> > just marked free and later refetched from the exec instead of from swap?o >>8 >> I'm surprised that any of them *don't* work that way. >> >sM >It's impossible to do it if the executable needs massaging in some way to bes >loaded. >tL >E.g. relocations on S/360 and its sucessors. As I recall, VM enabled you toL >build executables (called something else, but I recall not what) that couldJ >be used in that way but the price was that the (virtual) load address was >fixed at build time.   K They are called "shared segments" of various flavours.  They are not loadeddJ from a file system: they are effectively defined as pre-initialised pagingJ space.  CMS (component of VM/370 and descendants) uses them extensively.    L There is a way to build executables so they include no address constants, inJ which case relocation is not an issue, and the code can be executed at any8 load address without having to be modified at load time.  I Another approach is the TOC (Table Of Contents) approach used by AIX and  K others.  All address constants are kept together in a separate section, andmM the code segment contains no address constants, so can be loaded anywhere andi@ shared at different virtual addrs.  The TOC would not be shared.  L OS/360 and descendants (MVS, OS/390, z/OS) is a different story, one I'm notI very familiar with -- except for the basic object file format, wich otheroJ operating systems use too (from OS/390 on there is also an extended objectD file format, but relocatability issues remain pretty much the same).   Michel.b   ------------------------------  # Date: Fri, 19 Apr 2002 15:46:02 GMTn) From: Tom Van Vleck <thvv@multicians.org>   Subject: Re: Blade architecturesB Message-ID: <thvv-E7E15D.11453219042002@news.comcast.giganews.com>  ! hack@watson.ibm.com (hack) wrote:c  ( > John Homes <john.homes@eds.com> wrote:, > >"Peter da Silva" <peter@abbnm.com> wrote C > >> Charles Shannon Hendrix  <shannon@nospam.widomaker.com> wrote:aN > >> > How many systems out there try or have tried to do virtual memory whereK > >> > the executable file is part of mix, in a system where a code page iseO > >> > just marked free and later refetched from the exec instead of from swap?n > >>: > >> I'm surprised that any of them *don't* work that way. > >> > > O > >It's impossible to do it if the executable needs massaging in some way to be 
 > >loaded. > >oN > >E.g. relocations on S/360 and its sucessors. As I recall, VM enabled you toN > >build executables (called something else, but I recall not what) that couldL > >be used in that way but the price was that the (virtual) load address was > >fixed at build time.. > M > They are called "shared segments" of various flavours.  They are not loadedEL > from a file system: they are effectively defined as pre-initialised pagingL > space.  CMS (component of VM/370 and descendants) uses them extensively.   > N > There is a way to build executables so they include no address constants, inL > which case relocation is not an issue, and the code can be executed at any: > load address without having to be modified at load time. > K > Another approach is the TOC (Table Of Contents) approach used by AIX and  M > others.  All address constants are kept together in a separate section, andSO > the code segment contains no address constants, so can be loaded anywhere andLB > shared at different virtual addrs.  The TOC would not be shared.  6 Multicians, those of us who worked on Multics, used to8 hear stuff like this discussed, and look at each other,  and just make this gesture, of -" rubbing imaginary sticks together.  + The idea was, these folks are still rubbings, the rocks together, or rubbing the sticks on( the rocks, soon they'll figure out fire,. let's not even try to talk about electricity, , they'll just get mad.  But it makes you sad,- sometimes, seeing smart people re-invent the n- same good ideas, generation after generation.r  ( We called the TOC a "linkage section."  1 It was per-process, per-ring, and was initializeda6 by the dynamic linker when you first called a program.. Procedure code didn't need relocation and just, paged in as necessary, as Charles suggested.   ------------------------------   Date: 19 Apr 2002 16:22:51 GMT& From: peter@abbnm.com (Peter da Silva)  Subject: Re: Blade architectures- Message-ID: <a9pg8r$e9j@web.eng.baileynm.com>t  , In article <a9o2oh$8be$1@hermes.nz.eds.com>,& John Homes <john.homes@eds.com> wrote:5 > "Peter da Silva" <peter@abbnm.com> wrote in messaget) > news:a9o05n$bru@web.eng.baileynm.com...i; > > In article <slrnabpt5o.e4q.shannon@news.widomaker.com>,aB > > Charles Shannon Hendrix  <shannon@nospam.widomaker.com> wrote:M > > > How many systems out there try or have tried to do virtual memory whereEJ > > > the executable file is part of mix, in a system where a code page isN > > > just marked free and later refetched from the exec instead of from swap?  9 > > I'm surprised that any of them *don't* work that way.   N > It's impossible to do it if the executable needs massaging in some way to be	 > loaded.   L Ah, yes, if you don't have position independent code or enough address spaceK that you can map the shared libraries in at fixed offsets that's a problem.oL That probably also precludes paging the executable in as needed, you have to copy it in when it's executed.  L It would probably be worthwhile to allocate a chunk of storage to it (eitherF in swap or in a cache file associated with the executable) for paging,H if you can arrange things that the next time it's launched it's launchedK at the same address. PDP-11 UNIX did that for frequently executed programs:)F you would set what's called the "sticky bit", and after the first timeI they were executed their swap image would be left intact on the swap diskl? (which was often faster, and it was certainly simpler to load).s  I Microsoft's Pocket PC has a situation that's interesting: they may use VMnL to load executables from storage... but they do it even if the executable isI in the RAM disk in local memory, instead of executing it in place in RAM,aL because the RAM disk is compressed. Again, it would probably be worth havingL executables stored in expanded form and executed in place rather than havingN them paged from storage RAM to program RAM. With an access count and some kindI of garbage collector to recompress executables that are infrequently usedhK they could probably get a fairly significant savings in total memory usage.e   -- d+  `-_-'   In hoc signo hack, Peter da Silva.dE   'U`    "A well-rounded geek should be able to geek about anything." L                                                        -- nicolai@esperi.org          Disclaimer: WWFD?   ------------------------------   Date: 19 Apr 2002 16:25:48 GMT& From: peter@abbnm.com (Peter da Silva)  Subject: Re: Blade architectures- Message-ID: <a9pgec$ed3@web.eng.baileynm.com>e  0 In article <a9ojat$ee5$1@pegasus.csx.cam.ac.uk>,) Nick Maclaren <nmm1@cus.cam.ac.uk> wrote: B > None of the above.  Remember that I am not referring to personal > computers.  F No, actually, I don't remember that. In fact I said that I was talkingI about personal computers in the message you were responding to, where you D expressed interest in an OS that would fail more often from hardwareG failures than software problems. In that context, perhaps my comment isy less surprising, eh?  C > Now, I know that you were referring to personal computers, but myqB > assumption is that the same almost certainly holds true on them.C > Yes, the root case may be hardware, and the failure may look likeo@ > hardware, but the fact that the failure is so horrible is very > likely to be software.  + Such as the "software failures" I decribed?w   -- b+  `-_-'   In hoc signo hack, Peter da Silva.pE   'U`    "A well-rounded geek should be able to geek about anything."nL                                                        -- nicolai@esperi.org          Disclaimer: WWFD?   ------------------------------   Date: 19 Apr 2002 16:41:22 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren)  Subject: Re: Blade architectures0 Message-ID: <a9phbi$dpj$1@pegasus.csx.cam.ac.uk>  V In article <a9pgec$ed3@web.eng.baileynm.com>, peter@abbnm.com (Peter da Silva) writes:3 |> In article <a9ojat$ee5$1@pegasus.csx.cam.ac.uk>,,, |> Nick Maclaren <nmm1@cus.cam.ac.uk> wrote:E |> > None of the above.  Remember that I am not referring to personala |> > computers.7 |>  I |> No, actually, I don't remember that. In fact I said that I was talkingrL |> about personal computers in the message you were responding to, where youG |> expressed interest in an OS that would fail more often from hardwareoJ |> failures than software problems. In that context, perhaps my comment is |> less surprising, eh?v  @ I wasn't surprised, but I didn't make myself clear.  The trouble> is that the term "generic PC" is likely to mean a machine that? is not being used as a personal computer in many arenas!  I was.) thinking about that scenario - see below.r  F |> > Now, I know that you were referring to personal computers, but myE |> > assumption is that the same almost certainly holds true on them. F |> > Yes, the root case may be hardware, and the failure may look likeC |> > hardware, but the fact that the failure is so horrible is veryi |> > likely to be software.e |> s. |> Such as the "software failures" I decribed?  ? Yes.  The main niggle that I have is that you ARE likely to seed@ the same 'hardware' failure twice when you are using generic PCsB in a cluster.  And, of course, when you are doing anything unusual to X Windows :-)     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679t   ------------------------------  # Date: Fri, 19 Apr 2002 06:43:02 GMTx4 From: "Matt Muggeridge" <Matt.Muggeridge@compaq.com>$ Subject: Re: Buffered IO performance> Message-ID: <W%Ov8.25423$o66.76177@news-server.bigpond.net.au>  A > I am facing a performance drop when number of client connection-C > increase as the buffered IO increase.  I use 'mon sys' command to D > checked, I suspect that the server is buffered IO bounded (>8000).  K That's normal.  The question is what kind of decrease are you seeing?  Justc> how dramatic is it?  Which IP stack and version are you using?  G > Is there any system para. that I could setup such that I could reducen> > the buffered IO without changing architecture of the server?  G There's no magic button here.  Each time your application reads from or K writes to the sockect, it counts as buffered IO.  Minimising BUFIO requirestH you to minimise these events.  Can you write or read your data in larger chunks?   6 What else does your application do that affects BUFIO?  F > 3. standard C select call with timeout 1ms for new client connection > or conection drop.  K I'm not sure what you're trying to do with the 1ms timeout.  Does your fileDL descriptor mask include the listen socket as well?  In case it helps, take aL look at how the select call is used in the example ECHO server supplied with9 TCP/IP Services, TCPIP$EXAMPLES:TCPIP$ECHO_SERVER_PLUS.C.e  H > I suspect the high rate of buffred io is caused by the 1ms time out onE > client sockets (could be > 90).  I minimize the timeout time to 1msh    L Do you mean >90 client connections?  If so, and assuming your clients aren'tJ hammering the system with high message rates, then I'd expect a reasonablyI configured system to manage that effortlessly.  Your system is configured  adequately... right?   Matt.a --= -------------------------------------------------------------t OpenVMS TCP/IP Engineering Enterprise Computing Group Compaq Computer Corporations Gold Coast, AUSTRALIAl= -------------------------------------------------------------h    , "wing" <wingwong@witty.com> wrote in message7 news:873e96d6.0204180119.2340015c@posting.google.com...e > Hi,a > E > I am on performance tuning of a server which listen to a socket for H > client connection.  The server poll messages by a third party function8 > call to get messages to send to the connected clients. > A > I am facing a performance drop when number of client connection C > increase as the buffered IO increase.  I use 'mon sys' command tosD > checked, I suspect that the server is buffered IO bounded (>8000). >r > The main logic,e( > 1. poll messages by the 3rd party call3 > 2. send the polled messages to client thro socketcF > 3. standard C select call with timeout 1ms for new client connection > or conection drop. >oH > I suspect the high rate of buffred io is caused by the 1ms time out onE > client sockets (could be > 90).  I minimize the timeout time to 1mseE > because i want the server call the 3rd party call asap (this is thes > requirement).t >hG > Is there any system para. that I could setup such that I could reduce > > the buffered IO without changing architecture of the server? >p > Thanks in advance, >c > Wing   ------------------------------  % Date: Fri, 19 Apr 2002 12:39:10 +0100i( From: Nic Clews <sendspamhere@127.0.0.1>$ Subject: Re: Buffered IO performance) Message-ID: <3CC001DE.1D51E1A1@127.0.0.1>.   wing wrote:   A > I am facing a performance drop when number of client connectionoC > increase as the buffered IO increase.  I use 'mon sys' command tosD > checked, I suspect that the server is buffered IO bounded (>8000).  E I've experienced this. On an SMP system it also causes extremely highi MPSYNCH mode time.  C It was resolved because there was in fact an application problem. A.A little X window sat there doing the total wrong thing when it was?D allegedly idle. Even the Wintel application/platform was exonerated.  3 My advice would be to "consider the wider picture".o -- n( Regards, Nic Clews CSC Computer Sciences nclews at csc dot comt   ------------------------------    Date: 19 Apr 2002 08:20:36 -0700" From: ewilts@ewilts.org (Ed Wilts). Subject: Console scan profile for Gigaswitches< Message-ID: <995e39b6.0204190720.b885a8e@posting.google.com>  F Does anyone have PCM/VCS scan profiles for Gigaswitches that they'd beD willing to share?  If not, are the messages documented anywhere such that I can write my own?   Thanks,6	    .../Edh mailto:ewilts@ewilts.org   ------------------------------  % Date: Fri, 19 Apr 2002 09:44:45 +0200/- From: Jouk Jansen <joukj@hrem.stm.tudelft.nl>e) Subject: Re: CVS V1.11.2 on OpenVMS Error/3 Message-ID: <3CBFE70D.570DC9CE@hrem.stm.tudelft.nl>/   Stephen L. De Rudder wrote:  > 9 > I get the following error when using CVS on OpenVMS 7.2/ >  > $ cvs checkout openssl > cvs server: Updating openssl; > cvs checkout: move away openssl/CHANGES; it is in the wayp > C openssl/CHANGESn > C As I remember well, this happens because OpenVMS file names on ODS2.E disks are case-insensitive (always uppercase). The solution was, as Io, remember, to use update instead of checkout.7 I myself use another solution: I moved to an ODS5 disk.1                       Jouk   ------------------------------  % Date: Fri, 19 Apr 2002 08:17:42 -0400w2 From: rdeininger@mindspring.com (Robert Deininger)) Subject: Re: DEC 3000/400 VMS 6.2 problemeK Message-ID: <rdeininger-1904020817420001@1cust237.tnt1.nashua.nh.da.uu.net>   4 In article <3CBF73E7.9030109@dnv.com>, Arne Bergseth <Arne.Bergseth@dnv.com> wrote:  I >Seems your machine is running happily now, but from the information you sF >gave, there may be a possible problem if you are going to update the 
 >firmware. >----h ># >>: >MI >>: >I don't know what other information is needed, but here's what know:~' >>: >    133 MHz, 128MB, PMAG-DA video. K >>: >    CPU devstat = OK KN15-BA -V5.1-S749-I196-sV1.0-DECchip 21064  P3.0R >>E >The version number   sV1.0  indicates that your machine has not beenaG upgraded to accept firmware updates after 1995. On this newsgroup there I was a message on 12 Jul 95 from dshort@nova.wright.edu reporting problems0- with firmware update on a machine like yours:s  C Um, this info is a little out of date.  Let me (hopefully) clarify.   B CPU devstat = OK KN15-BA -V5.1-S749-I196-sV1.0-DECchip 21064  P3.0  J The "V5.1" in this string is the firmware revision.  5.1 is very old.  7.0J is current.  Firmware is contained in flash ROMs, and can be updated using the kits available on the web.  H The "sV1.0" is the revision level of the serial ROM (SROM).  The SROM isE not flashable.  It is a progammed part that's installed in a socket.  H sV1.0 is the "old" SROM.  I just opened up my 3000-400 and checked.  The label on my SROM is "309E7"   H >"If you're upgrading the firmware of a DEC3000/400 with the update diskC labeled may1995 ( or after ) check the rev #'s on your system PROM.o< >The relevant PROM is on the system board, the bottom board.I >It is near the front right of the board, below where the CD player is onmI my units. If the PROM is at rev 309E7, do not upgrade your firmware!  Theo PROM needs to be at rev 333E7."s >---- H >The part number is 23-333E7-00 and I got one from Compaq Service not so# long ago, and it was not expensive.c  G I have successfully upgraded my system from firmware V6.4 to V7.0 usingtF this "old" SROM.  Since it certainly didn't ship from the factory with7 V6.4, someone before me upgraded from Vancient to V6.4.A  E There was a bug in the sV1.0 SROM code, which caused certain firmwareoI updates to scramble the flash ROMs and render the system unuseable.  As IhG understand it, later firmware updates work around the SROM bug, so theya* can be used safely even with the old SROM.  J I've seen references to the SROM bug from old firmware release notes, withC a warning not to upgrade firmware with the old SROM installed.  TherH warning is NOT in the latest firmware release notes.  (Of course, anyone5 doing the update WILL read the release notes, right?)-  H Since the DEC 3000 systems don't have the floppy-based fail-safe loader,J scrambling the flash ROMs leaves the system dead.  (The existing code fromF the flash ROMs is needed to update the ROMs to a later revision.)  I'mI pretty sure the flash ROMs can be repaired via the mini-console, but I'velI never found the documentation for the mini-console commands, or the specs.J for the mini-console serial port.  If anyone knows this stuff, I'd like to hear from them.o     -- Robert>   ------------------------------  + Date: Fri, 19 Apr 2002 06:40:57 +0000 (UTC)s- From: "Ake Liss" <ake.liss@avestapolarit.com> 4 Subject: Emulate "show calls" in application loggingH Message-ID: <1d7d0334c2018f4bce6019f78575d076.20077@mygate.mailgate.org>   Hi,e  7 I'd like to be able to call (from a 3GL-pgm) a function   LIB$SIGNAL_CALL_STACK or similarC that gives the same information regarding modules, line numbers andlC calling order as a traceback-dump from a crash or SHOW CALLS in thes	 debugger.w  D I assume this RTL-routine does not exist, but perhaps someone of youB guys have created this kind of function and is willing to share...   TIAh  	 -ke Lissa BlueMaster Data      --  8 Posted via Mailgate.ORG Server - http://www.Mailgate.ORG   ------------------------------  # Date: Fri, 19 Apr 2002 10:46:05 GMT  From: system@SendSpamHere.ORG 8 Subject: Re: Emulate "show calls" in application logging0 Message-ID: <00A0CAF6.51F4B975@SendSpamHere.ORG>  x In article <1d7d0334c2018f4bce6019f78575d076.20077@mygate.mailgate.org>, "Ake Liss" <ake.liss@avestapolarit.com> writes: >Hi, >n8 >I'd like to be able to call (from a 3GL-pgm) a function! >LIB$SIGNAL_CALL_STACK or similaraD >that gives the same information regarding modules, line numbers andD >calling order as a traceback-dump from a crash or SHOW CALLS in the
 >debugger. > E >I assume this RTL-routine does not exist, but perhaps someone of youeC >guys have created this kind of function and is willing to share...  >s >TIA > 
 >-ke Liss >BlueMaster Data t >t >. >-- 9 >Posted via Mailgate.ORG Server - http://www.Mailgate.ORGh  I I have a callable traceback utility (TMESIS$CALLABLE_TRACEBACK) that will ? do just what you are asking.  If you're interested, contact me.B   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COMg            eJ   "And of course, I'm a genius, so people are naturally drawn to my fiery I   intellect.  Their admiration overwhelms their envy!" -- Calvin & Hobbesl   ------------------------------  # Date: Fri, 19 Apr 2002 06:35:40 GMTa* From: "Bill Todd" <billtodd@metrocast.net>) Subject: Re: Fearless VMS Prognostication0A Message-ID: <0VOv8.104807$K5.8831893@bin5.nnrp.aus1.giganews.com>:  < "Terry C. Shannon" <terryshannon@attbi.com> wrote in message5 news:gNMv8.34201$%s3.13539385@typhoon.ne.ipsvc.net...o >H >sJ > > 17% of what? There remains considerable debate over the current number > > of active VMS customers. >n > Indeed there does! >n > >c > > Compared to what?n > >w > > Over what time frame?r > I > Well, if there was sequential growth of 17 percent from 3FQ01 to 4FQ01,e thatE > would not mean a hell of a lot given the big hit CPQ (and virtuallyb everyone > else) took in 3FQ.  J IIRC there was a recent comment at openvms.org (or possibly tru64.org, butK that would seem a bit strange) to the effect that 17% of VMS customers were J new customers - it's possible that this is what the number meant (or not).   >e > >t) > > In what markets: US? European? Asian?a > >s2 > > New customers or existing? (What's the split?) > >c/ > > The number itself is virtually meaningless.h >sL > Correct. Details, we need details. And by the way, a quick glance at CPQ'sH > 1FQ02 results doesn't seem to indicate any dramatic growth in the High. > Performance Systems (VMS and Tru64) space...   Indeed.   E The first red flag in the 'Business Overview' section was that GlobalnL Services revenue (which has been pretty steadily on the increase for quite aA while, was the darling of the 'bold' new Compaq before the merger J announcement, and appears to be the darling of the 'new, Carly-improved HPJ way' cHomPaq now) was down 3% year over year (hence *not* just the 'normalK seasonal decline', unless someone changed the seasons between last year andtI this).  A great deal of services revenue is locked in by prior purchases,eI which means that new services revenue must have taken a noticeable hit if H overall services revenue is down.  And, of course, something like 90% ofK services revenue comes from the enterprise products (as was stated recentlytL in an article I've referred to at least once before but am too lazy to trackJ down again this time), so that's probably where most of the hits occurred.  A Next we see that Enterprise Computing segment revenue is down 13%yK sequentially.  'Seasonal' again (a post-Christmas-rush slump in NSK sales)?-L No:  the financial statement .pdf indicates that it was down almost 20% yearJ over year - which is not good considering that the dot-bomb cliff began inJ 2000Q4, was well under way in 2001Q1, and was thought to have bottomed outL before the beginning of 2002Q1.  True, the 'Access' segment revenue was downG almost as much year over year, but since that never contributes much to 3 profit anyway it's not clear why anyone would care.j  E And (jumping .pdfs yet again to the financial discussion) we see thateJ corporate consolidated gross margin dropped 2% year over year (from 23% toK 21% of revenue), due to "the competitive pricing environment and a shift inlD demand to low-cost technology alternatives" (could it be because theL high-cost alternatives were suddenly perceived as having been cut off at theL knees?).  Profitability was noticeably down year over year as well, once the" 'special items' were factored out.  J Sticking with the discussion .pdf, we see that in the Enterprise ComputingJ segment industry-standard server year over year revenue was hit the least:J while average prices dropped 18%, total units sold were up, hence the yearK over year revenue decrease was some number less than 18% (the press releasewH .pdf indicates that ISSG revenue was flat with 2001Q4, but it's too lateK tonight to chase down enough other data for other quarters to see whether anL better estimate of year over year figures can be made).  Storage revenue wasF down 19% year over year and 13% sequentially (funny how that expensive1 storage tends to track the expensive systems...).(  L Therefore, since overall Enterprise Computing revenue was down more than 19%K year over year, Alpha systems and NSK were, averaged, clearly down over 20%nK year over year.  Just how much is not made clear (and, again, it's too latecI to try to piece together something from past quarterly information) - buteL the press release .pdf states that BCSG revenue was down *33%* sequentially.  J Until now NSK has been the bulwark of BCSG revenue, sustaining it when allI other contributors declined, and the press release goes out of its way toeL state that "New customers comprised 16% of revenue for fault-tolerant [whichJ should mean NSK] systems".  This leads to at least a strong suspicion thatJ Alpha sales well and truly tanked, and that their apparent mini-revival inH Q4 last year was just what several of us suspected:  a blip due to a fewG large sales that likely were pretty much locked in before the Alphacideg
 announcement.n   - bill   ------------------------------  + Date: Fri, 19 Apr 2002 09:15:41 +0000 (UTC)y From: david20@alpha1.mdx.ac.uk) Subject: Re: Fearless VMS Prognostication>+ Message-ID: <a9on7t$nnl$1@aquila.mdx.ac.uk>e  o In article <gNMv8.34201$%s3.13539385@typhoon.ne.ipsvc.net>, "Terry C. Shannon" <terryshannon@attbi.com> writes:t >e > I >> 17% of what? There remains considerable debate over the current numberl >> of active VMS customers.  >e >Indeed there does!' >d >> >> Compared to what? >> >> Over what time frame? >uM >Well, if there was sequential growth of 17 percent from 3FQ01 to 4FQ01, that M >would not mean a hell of a lot given the big hit CPQ (and virtually everyonee >else) took in 3FQ.r >, >>( >> In what markets: US? European? Asian? >>1 >> New customers or existing? (What's the split?)t >>. >> The number itself is virtually meaningless. > K >Correct. Details, we need details. And by the way, a quick glance at CPQ'sjG >1FQ02 results doesn't seem to indicate any dramatic growth in the High-- >Performance Systems (VMS and Tru64) space...r >c >r  J According to statements made at the recent London OpenVMS Technical Update@ this is a 17% increase in new customers over the past 12 months.      
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  % Date: Fri, 19 Apr 2002 09:06:28 -0700c' From: David Mathog <mathog@caltech.edu>e) Subject: Re: Fearless VMS Prognosticationh+ Message-ID: <3CC04084.9EAC10C5@caltech.edu>    david20@alpha1.mdx.ac.uk wrote:? > L > According to statements made at the recent London OpenVMS Technical UpdateB > this is a 17% increase in new customers over the past 12 months.  8 Not nearly enough information is available to figure out; what's really going on. For all we know the 17% increase ino= new customers may all represent individuals buying a personalmB DS10. (This is exactly the sort of sales increase one would expect< whenever Alphas are blown out at firesale prices.)  The flip< side could be a 33% decrease in sales to existing customers,A each representing a much larger system not purchased.  That wouldeG be a revenue disaster but could still be spun with "17% new customers".C; Consider that they could truthfully make the "17% increase" @ statement if they had 6 new customers last year and 7 this year.   Regards,   David Mathog mathog@caltech.edu   ------------------------------  % Date: Fri, 19 Apr 2002 12:43:42 -0400a- From: JF Mezei <jfmezei.spamnot@videotron.ca>n) Subject: Re: Fearless VMS Prognostication , Message-ID: <3CC0493D.FEC0247A@videotron.ca>   david20@alpha1.mdx.ac.uk wrote: K > >> 17% of what? There remains considerable debate over the current numberp > >> of active VMS customers.p  K If they had 6 new customers the previous year, getting 7 new customers this-M year means a 17% increase. And that does not take into consideration how manygP existing customers stopped developping on VMS and how many actually dropped VMS.   ------------------------------  # Date: Fri, 19 Apr 2002 15:34:31 GMTo1 From: "Terry C. Shannon" <terryshannon@attbi.com>l) Subject: Re: Fearless VMS Prognosticationl; Message-ID: <bOWv8.34286$%s3.13925724@typhoon.ne.ipsvc.net>m  + <david20@alpha1.mdx.ac.uk> wrote in messaget% news:a9on7t$nnl$1@aquila.mdx.ac.uk...dG > In article <gNMv8.34201$%s3.13539385@typhoon.ne.ipsvc.net>, "Terry C. ) Shannon" <terryshannon@attbi.com> writes:  > >  > >lK > >> 17% of what? There remains considerable debate over the current numbere > >> of active VMS customers.s > >r > >Indeed there does!m > >1 > >> > >> Compared to what? > >> > >> Over what time frame? > >hJ > >Well, if there was sequential growth of 17 percent from 3FQ01 to 4FQ01, thatF > >would not mean a hell of a lot given the big hit CPQ (and virtually everyone > >else) took in 3FQ.p > >s > >>* > >> In what markets: US? European? Asian? > >>3 > >> New customers or existing? (What's the split?)h > >>0 > >> The number itself is virtually meaningless. > >pG > >Correct. Details, we need details. And by the way, a quick glance ata CPQ's,I > >1FQ02 results doesn't seem to indicate any dramatic growth in the Highj/ > >Performance Systems (VMS and Tru64) space...C > >. > >1 >1L > According to statements made at the recent London OpenVMS Technical UpdateB > this is a 17% increase in new customers over the past 12 months. >   F Would that there were a 17 percent increase in revenue! Anyhow, the 17I percent new customer figure tracks well with previous statistics. For theeL past four or five years the new customer proportion of the VMS sales mix has been 15 to 20 percent.  I Imagine what it would be like if more prospective new customers knew thatl the OS existed!W   ------------------------------  # Date: Fri, 19 Apr 2002 13:54:34 GMTi9 From: Hein van den Heuvel <hein_netscape@eps.zko.dec.com>u' Subject: Re: File IO query. FLAME ALERT-/ Message-ID: <3CC02002.4D54D355@eps.zko.dec.com>    JF Mezei wrote:m   > Hein van den Heuvel wrote:M > >   adds a records, it will hold the dirty buffer. In order for stream 2 to0M > >   add its record, RMS will have to tell stream 1 (through a blocking AST)sM > >   to write out that buffer, and then re-read into stream 2 private buffera > >   space. >tP > Would global buffers not seriously reduce actual disk io for the above ? Could > it conceptually do so ?a  Q Wouldn't help. During the original implementation of global buffers it was deemed Q too hard to have 'dirty' global buffers. The solution chosen back then, and still. activeQ now, is to copy the contents of a global buffer to a local buffer if modificationS isO requested. This is also the reason why certain buffers (SIDRs) are brought into & local buffers, not global, right away.  N The problem with dirty global buffers is that there would be no clear 'owning'K process. So if an other node in the cluster, or a process on the local nodeTP bypassing global buffers, were to cause a blocking ast that should force a flushK then then where does one deliver that. That problem could have been solved,-G and perhaps should have been solved, with a detached RMS server processrK but this has not been done. (Such server could also implement a 'lazy writeMJ behind' on deferred write buffers, and perhaps a read ahead and much more)   ------------------------------  % Date: Fri, 19 Apr 2002 12:49:04 -0400A- From: JF Mezei <jfmezei.spamnot@videotron.ca>h' Subject: Re: File IO query. FLAME ALERTi, Message-ID: <3CC04A7E.A144B854@videotron.ca>   Hein van den Heuvel wrote:M > process. So if an other node in the cluster, or a process on the local noderR > bypassing global buffers, were to cause a blocking ast that should force a flush) > then then where does one deliver that. 9  F Thanks. Makes sense. I guess in a non clustered system, one might haveO considered shared dirty buffers, but in a cluster, you couldn't really do that.5   ------------------------------   Date: 19 Apr 2002 04:25:20 GMT2 From: "Zane H. Healy" <healyzh@shell1.aracnet.com>6 Subject: Re: getting SIMH 2.9-6/VAX + VMS configured ?+ Message-ID: <a9o67g0400@enews2.newsguy.com>   + Brian Wheeler <bdwheele@indiana.edu> wrote:17 > well, you can point the emulator to the cdrom device:e   > set rq1 rrd40  > attach rq1 /dev/cdroms  = > and then boot the emulator and boot from dua1 (the real cd)d  L Have you gotten this to work?  I'd tried this with the PDP-11 emulator a fewJ months ago and it wouldn't work.  I'm assuming you're running Linux as the host OS?   		Zane  H PS just incase anyone asks, no I wasn't trying to boot VMS on the PDP-11 emulator :^)   ------------------------------  + Date: Fri, 19 Apr 2002 13:38:21 +0000 (UTC)i* From: bdwheele@indiana.edu (Brian Wheeler)6 Subject: Re: getting SIMH 2.9-6/VAX + VMS configured ?2 Message-ID: <a9p6kd$1lr$1@wilson.uits.indiana.edu>  + In article <a9o67g0400@enews2.newsguy.com>, 5 	"Zane H. Healy" <healyzh@shell1.aracnet.com> writes: - > Brian Wheeler <bdwheele@indiana.edu> wrote: 8 >> well, you can point the emulator to the cdrom device: >  >> set rq1 rrd40 >> attach rq1 /dev/cdrom > > >> and then boot the emulator and boot from dua1 (the real cd) > N > Have you gotten this to work?  I'd tried this with the PDP-11 emulator a fewL > months ago and it wouldn't work.  I'm assuming you're running Linux as the
 > host OS?  K Yes to both.  I did it yesterday & installed some layered products as well.w  ; Now if I could just get the DZ adapter to work under VMS.  u  / On simh 2.9-6 I set up the dz device like this:a   set dz lines=8 attach dz 10000e   and a "show dz" in simh shows:  H DZ, address=20000040-20000047, lines=8, attached to 10000, 0 connections    O and when I boot into VMS, the devices TTA0 -> TTA3 are available (why are there  only 4 devices?)  ' systartup_vms.com contains these lines:r    E $ SET TERM/PERM/MODEM/DIALUP/HANGUP/AUTOBAUD/ALTYPEAHD/HOSTSYNC TTA0: E $ SET TERM/PERM/MODEM/DIALUP/HANGUP/AUTOBAUD/ALTYPEAHD/HOSTSYNC TTA1:a2 $ SET SECURITY /CLASS=DEVICE /PROTECTION=W:R TTA0:2 $ SET SECURITY /CLASS=DEVICE /PROTECTION=W:R TTA1:  0 "show dev" in vms shows this for the TT devices:  . TTA0:                   Online               1. TTA1:                   Online               1. TTA2:                   Online               0. TTA3:                   Online               0  B [the error count for the two that have been set term'd is 1...odd]  : I can telnet to port 10000 but I never get a login prompt.   Any ideas what I'm doing wrong?e     Brianc       >  > 		Zane > J > PS just incase anyone asks, no I wasn't trying to boot VMS on the PDP-11 > emulator :^) >    ------------------------------   Date: 19 Apr 02 07:07:08 PST From: mckinneyj@cpva.saic.com_2 Subject: Re: How to prevent smtp relay w/Multinet?( Message-ID: <kO2ylmEibEaW@cpva.saic.com>  0 In article <a9ni0j$ej4$1@grapevine.wam.umd.edu>,-  bleau@umtof.umd.edu (Lawrence Bleau) writes:t  F See http://www.multinet.process.com/ftp/docs/html/admin_guide/Ch15.htmF for rejecting mail messages using a MULTINET:SMTP_SERVER_REJECT. file.    K > A colleague of mine is running OpenVMS VAX 6.2 and Multinet (I don't know-H > the version).  He recently received an email message saying his systemJ > has been added to a list of systems that allow open relay emails, and isI > therefore being blocked.  He went through all of hte Multinet menus andd5 > didn't see anything that looks like a relay option.g > I > I know what smtp relay is, and I closed it off on the systems I manage.bJ > I'm running UCX/TCPIP, however, so I can't give him the commands to use./ > I've never managed Multinet, so I've no hint.s > G > Is there a way to close off smtp relay in Multinet?  What is the menuAL > entry, or command if there is no menu option for it?  What is the earliestL > version of Multinet this command is available?  I'll ask my colleague whatK > Multinet version he's running - apologies, I know I should have done thiss > first.  Thanks.G >  > Lawrence Bleau > University of Maryland$ > Physics Dept., Space Physics Group > 301-405-6223 > bleau@umtof.umd.edut   ------------------------------    Date: 19 Apr 2002 13:56:04 -00005 From: "Doc.Cypher" <doc_cypher@redneck.gacracker.org>T2 Subject: Re: How to prevent smtp relay w/Multinet?5 Message-ID: <20020419135604.8326.qmail@gacracker.org>   5 NOTE: This message was sent thru a mail2news gateway.a8 No effort was made to verify the identity of the sender.8 --------------------------------------------------------  @ On Thu, 18 Apr 2002, bleau@umtof.umd.edu (Lawrence Bleau) wrote:J >A colleague of mine is running OpenVMS VAX 6.2 and Multinet (I don't knowG >the version).  He recently received an email message saying his system5I >has been added to a list of systems that allow open relay emails, and isiH >therefore being blocked.  He went through all of hte Multinet menus and4 >didn't see anything that looks like a relay option. > H >I know what smtp relay is, and I closed it off on the systems I manage.I >I'm running UCX/TCPIP, however, so I can't give him the commands to use.t. >I've never managed Multinet, so I've no hint. > F >Is there a way to close off smtp relay in Multinet?  What is the menuK >entry, or command if there is no menu option for it?  What is the earliest7K >version of Multinet this command is available?  I'll ask my colleague what J >Multinet version he's running - apologies, I know I should have done this >first.  Thanks.  J As others point out, you need to set up a SMTP_SERVER_REJECT file. There's$ a sample one in the documentation...  7 http://vmsbox.cjb.net/docs/admin_guide/Ch15.htm#E14E137C  I Of course, having found out by being blacklisted you'll also have the joyh# of trying to get off the blacklist.4     Doc. -- n6 The bigger the humbug, the better people will like it.K ~ Phineas Taylor Barnum.                             https://vmsbox.cjb.net    ------------------------------    Date: 19 Apr 2002 14:01:41 -0000= From: Doc.Cypher <Use-Author-Supplied-Address-Header@[127.1]>l2 Subject: Re: How to prevent smtp relay w/Multinet?5 Message-ID: <20020419140141.8552.qmail@gacracker.org>m  F On 19 Apr 2002, "Doc.Cypher" <doc_cypher@redneck.gacracker.org> wrote:   <snip>  G >>Is there a way to close off smtp relay in Multinet?  What is the menuhL >>entry, or command if there is no menu option for it?  What is the earliestL >>version of Multinet this command is available?  I'll ask my colleague whatK >>Multinet version he's running - apologies, I know I should have done thisr >>first.  Thanks.  >aK >As others point out, you need to set up a SMTP_SERVER_REJECT file. There'ss% >a sample one in the documentation...e > 8 >http://vmsbox.cjb.net/docs/admin_guide/Ch15.htm#E14E137 >wJ >Of course, having found out by being blacklisted you'll also have the joy$ >of trying to get off the blacklist.    > I forgot to add that Multinet questions can be directed to theA vmsnet.networks.tcp-ip.multinet newsgroup if your ISP carries it.a     Doc. -- y6 The bigger the humbug, the better people will like it.K ~ Phineas Taylor Barnum.                             https://vmsbox.cjb.nett   ------------------------------  # Date: Fri, 19 Apr 2002 16:03:23 GMTt7 From: tauberg@nauticom.no_spam_please.net (Jim Tauberg)m2 Subject: Installing Netscape Gold from Hobbyist CD1 Message-ID: <3cc03c89.54948565@news.nauticom.net>p   Hi,y7 	I can't figure out how to install the Netscape that iscC included on the VMS Hobbyist CD-ROM.  It appears to be there in twoc
 locations:  C One is from the root in a directory called [NETSCAPE_GOLD301.DIR;1]y% in this directory is one file called:e1 NETSCAPE-EXPORT-VAX-V303C4R-GOLD.EXE-DCX_VAXEXE;1o  5 This file also appears to be in the [KITS] directory.   B Since this file seems to work differently from the Layered Product= Installs, I haven't figured this out.  How do I install this?e   Thanks,i Jim Tauberg # tauberg@no_spam_please.nauticom.neta   ------------------------------    Date: 17 Apr 2002 18:07:11 -0700. From: David Gay <dgay@barnowl.CS.Berkeley.EDU> Subject: Re: Itanium troublesw5 Message-ID: <s71y9fln7b4.fsf@barnowl.CS.Berkeley.EDU>e  ( peter@abbnm.com (Peter da Silva) writes:  - > In article <a9e8qt$n1i5@news.kvaerner.com>,a6 > Tarjei T. Jensen <tarjei.jensen@kvaerner.com> wrote:F > > And this process also seem to ensure that there is few fundamentalN > > improvements. The shells have not changed much since I started using Unix.J > > About the only improvement is the general availability of command line > > editing. > N > Well, you know, the UNIX shell in 1981 and the UNIX shell in 2002 has a heckJ > of a lot more changes than the DOS/NT shell in 1981 and in 2002. In UNIXM > we have the addition of shell functions and aliases, job control, automaticnO > execution of different shells for different scripts, history and command lineiM > editing, command completion, event triggers, all kinds of stuff. The DOS/NTfL > shell has added the path to the default prompt, lost the ability to selectJ > the "switch" character, and you can use cursor keys instead of F1-F3 for > command line editing.   I Actually they've also managed to add filename completion, and "insert thewG pathname of the file dropped in the command window" in XP's shell (yes,h I was surprised...).   -- c	 David Gayr dgay@acm.org   ------------------------------   Date: 19 Apr 2002 08:18:34 CDT= From: wayne@tachysoft.xxx.564897.killspam.00c7 (Wayne Sewell)1G Subject: Re: learning how to use a computer (was: Re: Itanium troubles)7. Message-ID: <0cLEEHxc$$QM@tachxxsoftxxconsult>  [ In article <3CBE1AC5.213D5ECE@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:  > John Smith wrote:t >> r? >> "David J. Dachtera" <djesys.nospam@fsi.net> wrote in messagei$ >> news:3CBCDED5.95156971@fsi.net... >> >K >> > Even better: run a 10/100BaseT cable from the car to a home LAN hub inhF >> > the garage and run something on your Win, Linux/KDE/Gnome, Mac orH >> > whatever desktop and see exactly what the car is complaining about,J >> > order parts on-line and arrange delivery, ... up/down-load music, map* >> > data, etc. via your home network, ... >> >( >> > I can see a lot of possibilities... >> nI >> "Honestly officerthere was nothing I could do. I got a BSOD and my carn >> crashed.  > 3 > Makes you wonder about Billyware on warships, eh?t  ' "wonder" isn't the word.  Try "cringe".n   Waynes -- .O =============================================================================== M Wayne Sewell, Tachyon Software Consulting  (281)812-0738  wayne@tachysoft.xxx : http://www.tachysoft.xxx/www/tachyon.html and wayne.html  K change .xxx to .com in addresses above, assuming you are not a spambot  :-)0O ===============================================================================c= Society Lady:  Are you familiar with the Great Wall of China?@5        Curly:  No, but I know a big fence in Chicago!    ------------------------------   Date: 19 Apr 2002 14:19:43 GMT& From: peter@abbnm.com (Peter da Silva)G Subject: Re: learning how to use a computer (was: Re: Itanium troubles)h- Message-ID: <a9p91v$4ce@web.eng.baileynm.com>n  A In article <1rNv8.86296$3L2.7568088@bin7.nnrp.aus1.giganews.com>,o) Bill Todd <billtodd@metrocast.net> wrote:lG > The antecedent of 'it' above was the research, not the implementationrI > (though when the statement is taken in isolation from the discussion ofrM > which it was a part that's not at all clear).  The context of the statementvL > was to rebut the suggestion that command-oriented interfaces were superiorJ > to GUIs in facilitating the common kinds of things PC users do (and that. > Apple watched them do during said research).  D Both command-oriented (textually scriptable) interfaces and GUIs areE superior for different things in facilitating common things all kindsmD of users (PC users, UNIX gurus, handheld computer users, any kind of< user) do. Neither can fill the gaps left by the other, as is3 demonstrated by things like Hypercard and VBScript.r  ? A system that provides a strong GUI and a weak command-orientedpC interface is going to seem easier, but you run into a glass ceilingoC fairly quickly and third-party products can only go so far... sincen; without a standard API for applications to interface to theo@ command-oriented interface most applications can't be automated.  D On the other hand a system that provides a strong CLI and a weak GUIG faces the user with a fair amount of memorization to become proficient, E but once that's accomplished there's a lot of things that become easy0B that aren't really possible otherwise. On the other hand graphical programs need a GUI.   The trick is finding balance.n  F Meanwhile... last night I was trying to figure out how to import a fewG hundred lines of CSV into an Outlook address book when the fields don'trC have a 1:1 relationship to the ones Outlook wants, and one group ofo9 fields have meanings that are dependent on another group.h  D The solution there is going to involve writing a few lines of Awk toA rearrange things so Outlook's happy. There are scripting tools iniB Windows that will do the same thing, but VBscript is less oriented' towards that kind of text manipulation.u  G Anyway, what I'd really like would be the Windows GUI, a window manager H competitive with things like Windowmaker, and UNIX under the hood. MacOS X may just do the trick.   -- t+  `-_-'   In hoc signo hack, Peter da Silva.rE   'U`    "A well-rounded geek should be able to geek about anything."rL                                                        -- nicolai@esperi.org          Disclaimer: WWFD?   ------------------------------   Date: 19 Apr 2002 15:07:43 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren)G Subject: Re: learning how to use a computer (was: Re: Itanium troubles)n0 Message-ID: <a9pbrv$8gq$1@pegasus.csx.cam.ac.uk>  - In article <a9pa2v$5pc@web.eng.baileynm.com>,e( peter@abbnm.com (Peter da Silva) writes:3 |> In article <a9gml7$nd4$1@pegasus.csx.cam.ac.uk>, , |> Nick Maclaren <nmm1@cus.cam.ac.uk> wrote:E |> > Oh, critically.  There is a LOT of evidence that WYSIWYG editors,E |> > are faster for most people - there are some who can produce texts0 |> > without seeing it, but many more who can't. |> -H |> I'm increasingly coming to think that the best way to produce text isD |> through outliners or forms. Finishing needs WYSIWYG, but entry isH |> really improved by an interface that operates on the semantics of the2 |> text rather than the details of its appearance.  B I am one of the followers of Lawrence Sterne in this respect - the> layout is part of the information.  It is interesting that theA current separation between layout and content (as hyped by Latex) ( is a reversion to 17th century practice.     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679c   ------------------------------   Date: 19 Apr 2002 14:37:19 GMT& From: peter@abbnm.com (Peter da Silva)G Subject: Re: learning how to use a computer (was: Re: Itanium troubles) - Message-ID: <a9pa2v$5pc@web.eng.baileynm.com>f  0 In article <a9gml7$nd4$1@pegasus.csx.cam.ac.uk>,) Nick Maclaren <nmm1@cus.cam.ac.uk> wrote:tB > Oh, critically.  There is a LOT of evidence that WYSIWYG editorsB > are faster for most people - there are some who can produce text- > without seeing it, but many more who can't.S  E I'm increasingly coming to think that the best way to produce text islA through outliners or forms. Finishing needs WYSIWYG, but entry issE really improved by an interface that operates on the semantics of the / text rather than the details of its appearance.<  B > Interestingly enough, my evidence was that the users' preferenceA > was often for the GUI, even if they knew that it was slower.  Iu5 > notice that effect in myself, but can't explain it.e  " Recognition is easier than recall.  K One of the really nice things in the last version or so of Windows is smarthM drop-down alternative menus... it gives you the advantages of smart file name J completion in CLIs like TOPS and its collateral descendents (Kermit, tcsh,N IOS, etc) while taking on none of the disadvantages. In Windows 2000 or later,H if you're entering a file name in a field, you can type it in, but whileK you're typing the computer is looking for possible completions of the inputD< and presents them in a drop-down menu. So instead of typing:  1 	"C:\Program Files\Netscape\Program\Communicator"t   I type:d   	"C:\Pro$Net$P$Co$$"   (where $ is the down key)d  I In CLIs you have to hit a command key to get the expansion to happen, andeL often a couple of times because you haven't typed enough to produce a uniqueJ match. That's annoying enough that I only use it when I'm really unsure of the command.   -- j+  `-_-'   In hoc signo hack, Peter da Silva. E   'U`    "A well-rounded geek should be able to geek about anything." L                                                        -- nicolai@esperi.org          Disclaimer: WWFD?   ------------------------------  % Date: Fri, 19 Apr 2002 09:54:01 +0200u- From: Jouk Jansen <joukj@hrem.stm.tudelft.nl>s9 Subject: Re: linking with pthread library for C++ programc3 Message-ID: <3CBFE939.27AFE0EF@hrem.stm.tudelft.nl>w   Sammy wrote: >  > Hi:i > ? >   I am trying to link a c++ program with Pthread library. And < > everything was going wee until i notice that i need to useH > cxx/name=(short,as_is) options in my program in order to make some JNI > call. (that's another topic) > D >   Anyway, to make the long story short.  Here is a simple program. > 
 > Test.cpp- #define pthread_mutex_init PTHREAD_MUTEX_INITs- #define pthread_mutex_lock PTHREAD_MUTEX_LOCKd1 #define pthread_mutex_unlock PTHREAD_MUTEX_UNLOCKn3 #define pthread_mutex_destroy PTHREAD_MUTEX_DESTROY  >  > #include <stdio.h> > #include <pthread.h> > " > int main(int argc, char* argv[]) > {, >   >         pthread_mutex_t mutex; > + >         pthread_mutex_init(&mutex, NULL);T > % >         pthread_mutex_lock(&mutex);n > ' >         for (int i = 0; i++; i < 100)t >         {  >           printf("looping"); >         }A > ' >         pthread_mutex_unlock(&mutex);/ > ( >         pthread_mutex_destroy(&mutex); >  >         return 0;/ > }= > F Problem : you compile with /name=(as_is,short) and the thread-routines* are in the libraries with uppercase names.5 Solution : use the defines I included in your program   ? Note : The Compaq Porting Library contains a whole list of such 
 redefines.                    JoukT   ------------------------------  % Date: Fri, 19 Apr 2002 07:48:41 -0700o, From: "James Gessling" <jgessling@yahoo.com>9 Subject: Re: linking with pthread library for C++ programn4 Message-ID: <a9paoa$58lhp$1@ID-46415.news.dfncis.de>  L Yes, you can use #defines to solve this problem, but what a pain.  EverytimeA you use a new routine in your program you have to remember to addaK the define?  Shared libraries that I maintain, and I seem to be doing a lottJ of it, get upper and lowercase entry point names.  Keeps the C programmersG and BASIC programmers all happy.  To achieve this, in the symbol vector() each routine gets two lines, for example:m   account_open=PROCEDURE, -h& ACCOUNT_OPEN/account_open=PROCEDURE, -  " See the linker manual for details.  0 I wish Compaq would do this for there stuff too.   Jime  : "Jouk Jansen" <joukj@hrem.stm.tudelft.nl> wrote in message- news:3CBFE939.27AFE0EF@hrem.stm.tudelft.nl...y > Sammy wrote: > >d > > Hi:t > >yA > >   I am trying to link a c++ program with Pthread library. Ande> > > everything was going wee until i notice that i need to useJ > > cxx/name=(short,as_is) options in my program in order to make some JNI  > > call. (that's another topic) > >aF > >   Anyway, to make the long story short.  Here is a simple program. > >  > > Test.cpp/ > #define pthread_mutex_init PTHREAD_MUTEX_INITs/ > #define pthread_mutex_lock PTHREAD_MUTEX_LOCK 3 > #define pthread_mutex_unlock PTHREAD_MUTEX_UNLOCKo5 > #define pthread_mutex_destroy PTHREAD_MUTEX_DESTROYf > >i > > #include <stdio.h> > > #include <pthread.h> > >h$ > > int main(int argc, char* argv[]) > > {e > >k" > >         pthread_mutex_t mutex; > >h- > >         pthread_mutex_init(&mutex, NULL);a > >d' > >         pthread_mutex_lock(&mutex);  > >0) > >         for (int i = 0; i++; i < 100)-
 > >         {-  > >           printf("looping");
 > >         }D > >p) > >         pthread_mutex_unlock(&mutex);u > >:* > >         pthread_mutex_destroy(&mutex); > >g > >         return 0;3 > > }l > >aH > Problem : you compile with /name=(as_is,short) and the thread-routines, > are in the libraries with uppercase names.7 > Solution : use the defines I included in your program- >-A > Note : The Compaq Porting Library contains a whole list of sucht > redefines. >o >                  Jouk    ------------------------------  % Date: Fri, 19 Apr 2002 02:53:09 -0300 % From: "BCA" <noproblem5000@yajoo.com>b= Subject: Lleg? la oportunidad que esperabas... CAMBI? TU VIDAd9 Message-ID: <iss.1cd8.3cbfb0ca.8989d.1@mx2.west.saic.com>    <html>   <head>8 <meta http-equiv=3D"Content-Language" content=3D"es-ar">= <meta name=3D"GENERATOR" content=3D"Microsoft FrontPage 5.0">I< <meta name=3D"ProgId" content=3D"FrontPage.Editor.Document">L <meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dwindows-1= 252">  <title>Pagina nueva 1</title>n </head>?   <body>  L <p align=3D"center"><b><font face=3D"Verdana" size=3D"6" color=3D"#FF0000">= Lecto $ "RE"Grabadora LITE-ON</font></b></p>  L <p align=3D"center"><b><font face=3D"Verdana" size=3D"6" color=3D"#FF0000">= U$S 89.-=20s FINAL</font></b></p>  N <p align=3D"center"><b><font face=3D"Verdana" color=3D"#0000FF">POR TRES D=CD= AS O HASTA=20 AGOTAR STOCK</font></b></p>o  L <p align=3D"center"><b><font face=3D"Verdana" color=3D"#FF0000">BUSINESS CO= MPUTERSn </font></b></p>a  L <p align=3D"center"><b><font face=3D"Verdana" color=3D"#FF0000">011 4362 00= 49 / 3371</font></b></p>m  - <p align=3D"center"><b><font face=3D"Verdana"a8 color=3D"#FF0000">busicomputers@yahoo.com</font></b></p>  
 <p>&nbsp;</p>-  L <p><font size=3D"2" face=3D"Arial">Esta info es para armadores de m=E1quina= s.=20rJ Si no sos del gremio o no te interesa nuestra info: escr=EDbenos un e-mail@ vac=EDo a noproblem5000@yahoo.com e inmediatamente te daremos de baja.</font></p>   </body>e   ------------------------------  % Date: Fri, 19 Apr 2002 11:15:40 +0200t2 From: "Dr. Otto Titze" <titze@ikp.tu-darmstadt.de>6 Subject: Mails with long lines (> 255 bytes) truncated2 Message-ID: <3CBFE03C.FF21A8C@ikp.tu-darmstadt.de>  C I am still using MX Mailer (V4.1) on VAX and Alpha (VMS 6.2). From hD time to time I get more and more mail messages from Outlook senders E with very long lines in the body of the mail. These are truncated at C? 255 bytes. This is the same if I look at the mails under VMS ort from PC (after IUPOP3).d
 Questions:$ - is this a feature of my MX-mailer? - is it a problem of VMS mail?" - is there a workaround available?  D Someone told me at our last DECUS (Germany) that there is a logical , for MX available, but I did not find it yet.   Regardse Otto  E PS: Years ago a played around with the mailer under UCX. But I had sotB much problems there that I preferred to stay with MX. In principleC I could try again with new TCP/IP under VMS 7.2 if the problem does D not exist there. Nevertheless I like MX and would like to stay with " it if the problem could be solved.  ,  -------------------------------------------, | Dr. Otto Titze, Kernphysik TUD           |, | Schlossgartenstr. 9, D-64289 Darmstadt   |, | titze@ikp.tu-darmstadt.de                |, | Tel: +49(6151)16-2916,FAX:16-4321        |,  -------------------------------------------   ------------------------------  % Date: Fri, 19 Apr 2002 12:15:58 +0100e' From: Elliott Roper <elliott@yrl.co.uk>X: Subject: Re: Mails with long lines (> 255 bytes) truncated2 Message-ID: <190420021215582489%elliott@yrl.co.uk>  A In article <3CBFE03C.FF21A8C@ikp.tu-darmstadt.de>, Dr. Otto Titzew" <titze@ikp.tu-darmstadt.de> wrote:  E > I am still using MX Mailer (V4.1) on VAX and Alpha (VMS 6.2). From -F > time to time I get more and more mail messages from Outlook senders G > with very long lines in the body of the mail. These are truncated at lA > 255 bytes. This is the same if I look at the mails under VMS orv > from PC (after IUPOP3).< > Questions:& > - is this a feature of my MX-mailer? yesi  > - is it a problem of VMS mail? yesh$ > - is there a workaround available? yess&  In help of the latest (beta) of MX :-   SETh     LOCALn       /LONG_LINESa  F  Specifies that lines exceeding  the  MAIL-11  255  character  maximumE should  not be trucated during delivery to what appears to be a localiE user.  This setting is disabled by default.   Enabling  this  settinghF  could  cause mail delivery via VMS MAIL's DECnet support to fail when9 MX cannot determine that a recipient is a DECnet address.    Have not tried it yet.   ------------------------------  % Date: Fri, 19 Apr 2002 14:36:06 +0200m2 From: "Dr. Otto Titze" <titze@ikp.tu-darmstadt.de>: Subject: Re: Mails with long lines (> 255 bytes) truncated3 Message-ID: <3CC00F36.87DC5445@ikp.tu-darmstadt.de>    Thanks for the fast answer.    Elliott Roper schrieb:( >  In help of the latest (beta) of MX :- >  > SETf > 	 >   LOCALi >  >     /LONG_LINESa > ? It is obviously a feature of a newer Version of MX (not presenth in my MX V4.1) - hmmm...   Regards  Otto  y,  -------------------------------------------, | Dr. Otto Titze, Kernphysik TUD           |, | Schlossgartenstr. 9, D-64289 Darmstadt   |, | titze@ikp.tu-darmstadt.de                |, | Tel: +49(6151)16-2916,FAX:16-4321        |,  -------------------------------------------   ------------------------------  # Date: Fri, 19 Apr 2002 14:53:55 GMT ' From: Rick Dyson <Rick-Dyson@UIowa.EDU>": Subject: Re: Mails with long lines (> 255 bytes) truncated) Message-ID: <3CC02F83.D9BA3A63@UIowa.EDU>>   "Dr. Otto Titze" wrote:0 > D > I am still using MX Mailer (V4.1) on VAX and Alpha (VMS 6.2). FromE > time to time I get more and more mail messages from Outlook senders F > with very long lines in the body of the mail. These are truncated atA > 255 bytes. This is the same if I look at the mails under VMS ore > from PC (after IUPOP3).. > Questions:& > - is this a feature of my MX-mailer?  > - is it a problem of VMS mail?$ > - is there a workaround available?  ? 	Yes, I believe it was supported in the pre-commercial release.sE Though not via an internal command.  I think it was a logical name...   @ 	And, as luck would have it (not ususally for me this week!), I 6 still have it commented out in my SyLogicals.com file:  ? $ Define /System /Executive_Mode MX_Local_Allow_Long_Lines Truet   Give that a try!   rick   ------------------------------  % Date: Fri, 19 Apr 2002 11:19:24 -0400-; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com>=: Subject: Re: Mails with long lines (> 255 bytes) truncated$ Message-ID: <3cc035dd$1@news.si.com>  C >I am still using MX Mailer (V4.1) on VAX and Alpha (VMS 6.2). From-D >time to time I get more and more mail messages from Outlook sendersE >with very long lines in the body of the mail. These are truncated atu
 >255 bytes  K First, post messages about MX in vmsnet.mail.mx.  Second, upgrade to MX 4.2." (the last freeware version of MX).  L MX V4.x does not have the ability to allow lines longer than 255 characters.G You need MX V5.1A or better for that.  MX V5.x is a commercial product.   E You can always tell your Outlook people to DO THE RIGHT THING and putn ends-of-line in their messages.  --A Brian Tillman                   Internet: tillman_brian at si.comeA Smiths Aerospace                          tillman at swdev.si.com = 3290 Patterson Ave. SE, MS      Addresses modified to preventi< Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  # Date: Fri, 19 Apr 2002 15:49:31 GMTd- From: goathunter@goatley.com (Hunter Goatley)t: Subject: Re: Mails with long lines (> 255 bytes) truncated0 Message-ID: <3cc03b15.89450262@news.process.com>  3 On Fri, 19 Apr 2002 11:19:24 -0400, "Brian Tillman"e, <tillman_brian@notnoone.notnohow.com> wrote:  D >>I am still using MX Mailer (V4.1) on VAX and Alpha (VMS 6.2). FromE >>time to time I get more and more mail messages from Outlook senderstF >>with very long lines in the body of the mail. These are truncated at >>255 bytesn >eL >First, post messages about MX in vmsnet.mail.mx.  Second, upgrade to MX 4.2# >(the last freeware version of MX).o >cM >MX V4.x does not have the ability to allow lines longer than 255 characters.o  F Not true.  There was a patch released for V4.2 that would allow longerI lines.  If memory serves, Matt took that out of MX V5.0, added it back inlK with the logical in V5.1, and then converted it to the SET LOCAL command inV V5.2.n  A You can find the V4.2 patch in the [.MX.MX042.PATCH] directory on D ftp.madgoat.com.  The source for that patch can be found in the file% MX042_SRC_UPDATE_WITH_ANTI_RELAY.ZIP.h   Hunter ------9 Hunter Goatley, Process Software, http://www.process.com/ 8 goathunter@goatley.com    http://www.goatley.com/hunter/< New Robert R. McCammon site: http://www.RobertRMcCammon.com/   ------------------------------  % Date: Fri, 19 Apr 2002 08:16:41 +0200i- From: Didier Morandi <Didier.Morandi@Free.fr> K Subject: Re: Major Virus Alert!  Better get on VMS or APPLE now or bye bye!a' Message-ID: <3CBFB649.4D38F16A@Free.fr>   D Why? I could have written the same for Macintosh/MacOS systems. It's@ just that VMS and MacOS are protected by the excellence their SW engineers put into their work.  H I spent since the beginning of the week more that 16 hours to remove theE new KLEZ virus from two PCs. I got that messages which propagated theeF virus, but as I use a Mac (I am an IT professional, you see :-) I just throwed the message away.g  % For you information, the message was:<  	 Subject: bK Date:  Mon, 15 Apr 2002 22:55:09 +0200 (added by postmaster@libertysurf.fr) U Sender: fred.toto@libertysurf.fr (derived from envelope by postmaster@libertysurf.fr)>   WIN_$100_NOW.DOC.pif   Name: WIN_$100_NOW.DOC.pif1 Type: unspecified type (application/octet-stream)h Encoding: base64  H This new virus is made of two files, PE_ELKERN and KLEZ.x (I had .E) ForC W98, the first one hides itself in KERNEL32.EXE and detects at eachsF startup (even without the starting applications disabled via msconfig)B if KLEZ is present. If you successfully succeeded to remove it viaF FIX_KLEZ.COM, it reinstalls it. You have to build and use first that 6D floppy disks (more than two hours) to eradicate the thing (describedF below). It also uses an ie "bug" which allows it to infect the PC even if you do not open any email attachment (see http://www.microsoft.com/technet/treeview/default.asp?url=/technet/security/bulletin/MS02-015.asp   The solution is at:uC http://www.antivirus.com/vinfo/security/readme_worm_klez.g_3.10.txt1   D.   "Mark(unMASK)Forsyth" wrote: > 7 > "Bob Ceculski" <bob@instantwhip.com> wrote in messagea9 > news:d7791aa1.0204180444.7e0cafd3@posting.google.com...dD > > this is from www.nwfusion.com ... looks like viruses are gettingC > > ready to explode every unix/linux/windoze box around, and theretC > > will not be an easy way to stop them ... unless you are on vms!a > >u > I > Who the hell are you, Bob ? Are you some sort of magician ? You must be K > getting close to it to be able to infer what you wrote, above,  from what  > you included, below.   -- n     Pascale et Didier Morandi   "Svoukyavachtlamaisondleroi"      19 chemin de la Butte         31400 Toulouse         06 79 83 64 18   ------------------------------    Date: 19 Apr 2002 08:47:08 -0700& From: jordan@ccs4vms.com (Rich Jordan)< Subject: Re: MicroVAX 3100-30 self test failure: RT-Retrbctr< Message-ID: <cc5619f2.0204190747.ff1eae9@posting.google.com>  E I guess maybe this is either an oddball problem or a genuine hardware E fault.  Further testing has been performed.  With the system stripped E to minimum (onboard memory only, no SCSI devices, AUI net on loopbacklD plug, DSW41 removed) it will gave repeated RT-Retrbctr errors duringD self test.  Interestingly enough, after a number of power cycles, itC suddenly passed self test.  I reinstalled all the options, one at a0B time, and each time the self test passed.  The only wierdness came@ during the first power up after reinstalling the DSW41, where itE passed, but doing a SHOW CONFIG would hang after displaying device 10SF SCSI as passing (device 12 would be the DSW41).  Next power cycle that cleared up.d  @ Assuming there's something about either the power cycles or justE having power applied for some time clearing the faults, I removed the C DSW41, left everything else on, and shut down overnight.  This AM IeE got the RT-Retrbctr errors again.  After power cycling about 8 times,rF it stopped occurring, and self test passed several consecutive cycles.  A There's no obvious battery or battery pack on this system.  Where.B would the battery be, and is it possible this is the fault?  MaybeE being powered on for a while gives a minimal charge that is enough to-= survive a 30 second power down, but not enough for overnight?   E Is this a cheaply repaired  problem (assuming that the battery is theoF problem)?  Keeping in mind that this system is bound for surplusing or, the dumpster if it can't be easily repaired.   Thanks for any info...   Rich Jordany    k jordan@ccs4vms.com (Rich Jordan) wrote in message news:<cc5619f2.0204170713.14f6b6ff@posting.google.com>...rD > We're prepping some old MicroVAX 3100s for surplus sale (maybe viaG > Ebay, not yet determined, but at least they're not getting pitched). eG > The one -30 I have worked OK yesterday; passed self test, booted, anduH > ran well enough for me to clean off the data disks and test the TZ30. G > This morning, it fails during the selftest (fboot set for fast test);uH > while drawing the progress bar on the terminal, it suddenly prints '??E > RT-Retrbtcr', hangs for about 20-30 seconds, then restarts the selfu& > test.  It repeats this continuously.
 > ........ > 
 > Rich Jordan    ------------------------------  % Date: Fri, 19 Apr 2002 10:25:39 -0700f# From: "Tom Linden" <tom@kednos.com>  Subject: Mime hole virus9 Message-ID: <LHENJCHGEGBPHFJFFKNFKEAICAAA.tom@kednos.com>   A I discovered what this is the hard way, hope I didn't pass it on.i> I run tcpip5.1 under 7.3 smtp server and use Outlook on W2K asH pop client.  If you have the preview turned on, TURN IT OFF,  apparentlyB this is all that is needed to launch it, and it comes with its own smtp server!  D I switched to a virgin backup system so I am not contagious anymore.   Tom0   ------------------------------  % Date: Fri, 19 Apr 2002 07:53:59 -0700l, From: "James Gessling" <jgessling@yahoo.com>2 Subject: OpenVMS E7.3-1 changes show dev d display4 Message-ID: <a9pb28$5569u$1@ID-46415.news.dfncis.de>  L Device                  Device           Error    Volume         Free  Trans MntrL  Name                   Status           Count     Label         Bytes Count CntoL TECH03$DKA0:            Mounted              0  ALPHASYS      821.77MB   383 1nL TECH03$DKA100:          Mounted              0  DATA            3.99GB     1 1i  I For example.  I kind of like the MB and GB free space, but it might breakaI some scripts people have cooked up for monitoring.  Sho dev dk/full looksm to be as it was before.r   Jim    ------------------------------  % Date: Fri, 19 Apr 2002 12:33:13 -0400 2 From: norm lastovica <norman.lastovica@oracle.com>6 Subject: Re: OpenVMS E7.3-1 changes show dev d display* Message-ID: <3CC046C9.AEF2E59E@oracle.com>  3 I understand that there is, or will be, a switch ore4 logical or something to restore the old behaviour as well.    James Gessling wrote:  > N > Device                  Device           Error    Volume         Free  Trans > MntaN >  Name                   Status           Count     Label         Bytes Count > Cnt N > TECH03$DKA0:            Mounted              0  ALPHASYS      821.77MB   383 > 1EN > TECH03$DKA100:          Mounted              0  DATA            3.99GB     1 > 1i > K > For example.  I kind of like the MB and GB free space, but it might breakgK > some scripts people have cooked up for monitoring.  Sho dev dk/full looksr > to be as it was before.) >  > Jim    --  > norman lastovica / oracle rdb engineering / usa / 610.696.4685   ------------------------------  # Date: Fri, 19 Apr 2002 16:43:50 GMT / From: brooks@cuebid.zko.dec.nospam (Rob Brooks)&6 Subject: Re: OpenVMS E7.3-1 changes show dev d display- Message-ID: <1QDrUxZ2hfky@cuebid.zko.dec.com>   c In article <a9pb28$5569u$1@ID-46415.news.dfncis.de>, "James Gessling" <jgessling@yahoo.com> writes:pP >Device                  Device           Error    Volume         Free  TransMntP >Name                   Status           Count     Label         Bytes  CountCntN > TECH03$DKA0:            Mounted              0  ALPHASYS      821.77MB   383 > 10N > TECH03$DKA100:          Mounted              0  DATA            3.99GB     1 > 1  > K > For example.  I kind of like the MB and GB free space, but it might break.K > some scripts people have cooked up for monitoring.  Sho dev dk/full looks= > to be as it was before.E >  > Jimn  H This will change for the SSB release.  The default behaviour will revertI to pre- E7.3-1.  There will be a new qualifier -- /SIZE that will controllI the display.  The default will be /SIZE=BLOCKS; /SIZE=BYTES will give you ( the display you get by default in E7.3-1  G I *think* there is a logical name that controls the display for E7.3-1.<  H Fortunately, that logical name will be replaced with the addition of the above-named qualifier.   -- sM Rob Brooks    VMS Engineering -- I/O Exec Group     brooks!cuebid.zko.dec.comD   ------------------------------  % Date: Fri, 19 Apr 2002 14:39:10 +0100p- From: Roy Omond <Roy.Omond@BlueBubble.UK.Com>oE Subject: Patch *entitlement*: what does *entitlement* actually mean ?e1 Message-ID: <3CC01DFE.EA9BE12B@BlueBubble.UK.Com>1   Gentle colleagues,  H I am in a bit of a quandary at the moment, and would appreciate comments  	 and help.i  G I do VMS software support for a small customer in the UK.  The customer E is particularly interested in the X25ALP X25_V0105-1 Compaq X.25 V1.5 A for OpenVMS patch.  This falls in the category of "Entitled Kits"l= cf. "Public Kits" (clear) and "Restricted Kits" (also clear).   E See http://ftp.support.compaq.com/patches/.new/openvms.shtml#entitlede  9 The customer has a VMS Media Update contract with Compaq.i  F Question:  am I legally entitled to download the patch kit from CompaqF on behalf of my customer ?  Is there any (preferably) on-line document> where this is described.  Please remember, I am not a lawyer !   many thanks in advance,e  	 Roy Omonds Blue Bubble Ltd.   ------------------------------  # Date: Fri, 19 Apr 2002 06:18:32 GMTt' From: "David G. Conroy" <dgc@spies.com> 2 Subject: Re: Predictions - just for the hell of itB Message-ID: <YEOv8.5077$543.3217411144@newssvr21.news.prodigy.com>  ; I'm not sure that this is correct. I vaguely remember beingoG interviewed by our (DEC's) lawyer when these suits were active, and I'mb? almost positive I remember him saying that the most troublesome 6 patents came from a cancelled VAX (ANDROMEDA perhaps?)  
 ----------; In article <qh8z7kfe3k.fsf@ruckus.brouhaha.com>, Eric Smithl) <eric-no-spam-for-me@brouhaha.com> wrote:f    0 > Paul Repacholi <prep@prep.synonet.com> writes:E >> 10 years later, it is the KC-10 patents that Intel have stolen for 1 >> the PPro. And the MESI cache protocol I think.r >t! > Got patent numbers (or titles)?|   ------------------------------  ! Date: Fri, 19 Apr 02 07:02:57 GMTt From: jmfbahciv@aol.come2 Subject: Re: Predictions - just for the hell of it+ Message-ID: <a9orf1$gs2$1@bob.news.rcn.net>-  , In article <a9k7op$4pnn$1@lead.zk3.dec.com>,9    "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote:  >j >Holy crap.s >tI >Pat?  Is that you?  Where've you been?  It's been a *long* time.  What'st >new?o >> > C >Patrick Sweeney wrote in message <3CBCCD86.539EDBB9@nyc.rr.com>...t <snip>  > An e-mail friend of mine was also glad to see him make a post.   /BAH  ' Subtract a hundred and four for e-mail.g   ------------------------------  # Date: Fri, 19 Apr 2002 09:40:17 GMTn' From: "Hans Vlems" <hvlems@zfree.co.nz>n Subject: Simh V2.9 problem" Message-ID: <3cbfe5f2@zfree.co.nz>  P Compiling sim V2.9 for the PDP-11 and the NOVA on a VAX/VMS V7.2 system resulted in an error.  @ Module sim_sock.h contains an #include <stropts.h> statement and that file was not found.B I created an empty container file and the file compiled all right.B Linking both the PDP-11 and NOVA executables resulted in warnings.   Compiler info: $ cc/ver" DEC C V5.6-003 on OpenVMS VAX V7.2 $r  ' Output of the @build_vms PDP11 process:s   $ @build_vms pdp11  + Runtime Debugger Won't Be Included At Link.v Compiling On A VAX Machine.p  2 Compling The SYS$DISK:[.LIB]PDP11-VAX.OLB Library.P Using Compile Command: CC/PREFIX=ALL/NOOPTIMIZE/NODEBUG/NEST=PRIMARY/NAME=(AS_IS) )/INCLUDE=(SYS$DISK:[],SYS$DISK:[.PDP11])   $         SYS$DISK:[.PDP11]PDP11_CIS.C$         SYS$DISK:[.PDP11]PDP11_CPU.C#         SYS$DISK:[.PDP11]PDP11_DZ.Cs#         SYS$DISK:[.PDP11]PDP11_FP.Cm#         SYS$DISK:[.PDP11]PDP11_IO.Cw#         SYS$DISK:[.PDP11]PDP11_LP.C #         SYS$DISK:[.PDP11]PDP11_RK.C #         SYS$DISK:[.PDP11]PDP11_RL.CL#         SYS$DISK:[.PDP11]PDP11_RP.Ct#         SYS$DISK:[.PDP11]PDP11_RQ.Cs#         SYS$DISK:[.PDP11]PDP11_RX.Ca'         SYS$DISK:[.PDP11]PDP11_STDDEV.C-$         SYS$DISK:[.PDP11]PDP11_SYS.C#         SYS$DISK:[.PDP11]PDP11_TC.Cg#         SYS$DISK:[.PDP11]PDP11_TM.C #         SYS$DISK:[.PDP11]PDP11_TS.C     % Building SYS$DISK:[.BIN]PDP11-VAX.EXE $ %LINK-W-WRNERS, compilation warningsA         in module SCP_TTY file DISK$DISK:[SIMH.LIB]SIMH-VAX.OLB;1p' %LINK-W-NUDFSYMS, 15 undefined symbols:i$ %LINK-I-UDFSYM,         tmxr_attach ! %LINK-I-UDFSYM,         tmxr_dep u$ %LINK-I-UDFSYM,         tmxr_detach # %LINK-I-UDFSYM,         tmxr_dscln e  %LINK-I-UDFSYM,         tmxr_ex $ %LINK-I-UDFSYM,         tmxr_fconns $ %LINK-I-UDFSYM,         tmxr_fstats % %LINK-I-UDFSYM,         tmxr_getc_ln  ! %LINK-I-UDFSYM,         tmxr_msg o' %LINK-I-UDFSYM,         tmxr_poll_conn  % %LINK-I-UDFSYM,         tmxr_poll_rx e% %LINK-I-UDFSYM,         tmxr_poll_tx I% %LINK-I-UDFSYM,         tmxr_putc_ln e& %LINK-I-UDFSYM,         tmxr_reset_ln " %LINK-I-UDFSYM,         tmxr_rqln 5 %LINK-W-USEUNDEF, undefined symbol tmxr_ex referencedi)         in psect dz_dev offset %X00000028vC         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1s6 %LINK-W-USEUNDEF, undefined symbol tmxr_dep referenced)         in psect dz_dev offset %X0000002CaC         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1L8 %LINK-W-USEUNDEF, undefined symbol tmxr_dscln referenced)         in psect dz_mod offset %X0000002C:C         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1c: %LINK-W-USEUNDEF, undefined symbol tmxr_poll_rx referenced(         in psect $CODE offset %X000002A9C         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1o: %LINK-W-USEUNDEF, undefined symbol tmxr_poll_rx referenced(         in psect $CODE offset %X00000585C         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1 6 %LINK-W-USEUNDEF, undefined symbol tmxr_msg referenced(         in psect $CODE offset %X000006B3C         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;15; %LINK-W-USEUNDEF, undefined symbol tmxr_reset_ln referencedi(         in psect $CODE offset %X000006BEC         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1e: %LINK-W-USEUNDEF, undefined symbol tmxr_poll_tx referenced(         in psect $CODE offset %X00000707C         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1e: %LINK-W-USEUNDEF, undefined symbol tmxr_putc_ln referenced(         in psect $CODE offset %X000007ADC         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1y: %LINK-W-USEUNDEF, undefined symbol tmxr_poll_tx referenced(         in psect $CODE offset %X000007B7C         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1)< %LINK-W-USEUNDEF, undefined symbol tmxr_poll_conn referenced(         in psect $CODE offset %X0000083BC         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1s: %LINK-W-USEUNDEF, undefined symbol tmxr_poll_rx referenced(         in psect $CODE offset %X000008BAC         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1 : %LINK-W-USEUNDEF, undefined symbol tmxr_poll_tx referenced(         in psect $CODE offset %X000008CBC         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1 : %LINK-W-USEUNDEF, undefined symbol tmxr_getc_ln referenced(         in psect $CODE offset %X00000934C         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1 7 %LINK-W-USEUNDEF, undefined symbol tmxr_rqln referenced (         in psect $CODE offset %X000009E4C         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1a9 %LINK-W-USEUNDEF, undefined symbol tmxr_attach referenced.(         in psect $CODE offset %X0000103FC         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1 9 %LINK-W-USEUNDEF, undefined symbol tmxr_detach referenced (         in psect $CODE offset %X000010F2C         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1o9 %LINK-W-USEUNDEF, undefined symbol tmxr_fconns referencedp(         in psect $CODE offset %X000011E9C         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1 9 %LINK-W-USEUNDEF, undefined symbol tmxr_fstats referenced.(         in psect $CODE offset %X00001207C         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1a6 %LINK-W-USEUNDEF, undefined symbol tmxr_msg referenced(         in psect $CODE offset %X00001344C         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1s; %LINK-W-USEUNDEF, undefined symbol tmxr_reset_ln referencedu(         in psect $CODE offset %X00001358C         in module PDP11_DZ file DISK$DISK:[SIMH.LIB]PDP11-VAX.OLB;1m $   I The problem is very likely the missing stropts.h file. Any ideas where tod get it?s       http://www.zfree.co.nz    .   ------------------------------  # Date: Fri, 19 Apr 2002 14:50:36 GMTr' From: Rick Dyson <Rick-Dyson@UIowa.EDU>1  Subject: Re: SSH for Alpha OVMS?) Message-ID: <3CC02EBC.4BE45DDF@UIowa.EDU>    Kiasu Surfer wrote:sE > But OpenVMS, choices are limited. We don't intend to pay extras forrE > these security enhancement, and we are further handicapped with the + > lack of compilers, namely the C compiler.a< > We are unable to compile the available source found at OSU2 > ( http://kcgl1.eng.ohio-state.edu/~jonesd/ssh/ )  G What was the problem building the OSU SSH v1.5 kit?  I have built it on F VAXen and Alpha from OpenVMS v6.2 -> v7.3 without a hitch.  As well as% the required OpenSSL (up to v0.9.6a).d  E Until the Compaq TCP/IP has a built-in SSH server that works reliablyeF (note there is none yet and when I asked at a teleconference last yearF they said they had no immediate plans, so if it is being reported thatJ it is coming, I am curious how well it is put together in such a hurry...)   Rick   ------------------------------  % Date: Fri, 19 Apr 2002 10:32:03 +0100g4 From: John Laird <john@laird-towers.freeserve.co.uk>A Subject: Re: Stuck in licenseless Motif with OpenVMS 7.2 Hobbyisti8 Message-ID: <9qmvbuk72lm4eq1dj5afn8fhpk47qh8l2g@4ax.com>  K On Fri, 19 Apr 2002 00:38:42 -0400, JF Mezei <jfmezei.spamnot@videotron.ca>  wrote:   >Jim Tauberg wrote: F >> installed MOTIF and after the install was asked to reboot.  Now theE >> system comes up in a graphical format, but when I try to log in asdE >> SYSTEM, it shows a dialog that says "No license is active for thist >> software product" >e> >Press the reset button or break key to get to the >>> prompt. >  >Then B/1 to get to SYSBOOTr >   >at SYSBOOT> SET WINDOW_SYSTEM 0 >    SYSBOOT> CONTINUE >3K >This will bring up your machine without any windowing and just a very verysM >basic character cell interface (not even VT100, more like TTY). You can theneF >type in the licence information, then use SYSGEN to USE CURRENT , SET1 >WINDOW_SYSTEM 1 , WRITE CURRENT and then reboot.I  I Decwindows has a habit of just starting anyway, ime.  (If you look at the-L startup procedure, you'll see it assumes that WINDOW_SYSTEM set to zero must mean 1 of course.)  L Probably easier to do a minimum boot, although this may get you into trouble; if you rely too much on logical names that are not defined:t   SYSBOOT> SET STARTUP_P1 "MIN"i SYSBOOT> SET WRITESYSPARAMS 0 
 SYSBOOT> CONTW  I The second line will prevent an implicit WRITE CURRENT, so you can simplyu reboot normally afterwards.      	John    ------------------------------    Date: 19 Apr 2002 06:56:47 -0700- From: jodonnell@hrblock.com (Jason O'Donnell)cA Subject: Re: Stuck in licenseless Motif with OpenVMS 7.2 Hobbyist = Message-ID: <9059bf6b.0204190556.431f80e7@posting.google.com>s  o tauberg@nauticom.no_spam_please.net (Jim Tauberg) wrote in message news:<3cbf7146.2906040@news.nauticom.net>.... > Hi,r@ > 	I have a VAXstation 4000 VLC which I've installed the OpenVMSB > 7.2 hobbyist on.  I'm very much a beginner with this.  Anyway, IE > installed MOTIF and after the install was asked to reboot.  Now theuD > system comes up in a graphical format, but when I try to log in asD > SYSTEM, it shows a dialog that says "No license is active for thisG > software product" and gives me an "OK" button which returns me to theuG > graphical login screen.  Of course it's not licensed, cause I haven'tJH > put in the license information yet... but I don't know how to get backF > to the $ prompt to do it.   Any help would be appreciated!  (I don't5 > even know how to safely shut it down at this point)o >  > Thanks in advance,
 > Jim Taubergt > tauberg@no_spam.nauticom.net  B You may also get layered product licenses at the same site you gotF your hobbyist license.  It includes motif licenses.  Heck you can even get ALL-IN-1. ;)   ------------------------------  # Date: Fri, 19 Apr 2002 15:26:11 GMTr7 From: tauberg@nauticom.no_spam_please.net (Jim Tauberg)RA Subject: Re: Stuck in licenseless Motif with OpenVMS 7.2 Hobbyist 1 Message-ID: <3cc036f7.53522883@news.nauticom.net>    Hi,h- 	Thanks for the help.  This worked perfectly.a
 Sincerely, Jim T.      . On Fri, 19 Apr 2002 10:32:03 +0100, John Laird* <john@laird-towers.freeserve.co.uk> wrote:  L >On Fri, 19 Apr 2002 00:38:42 -0400, JF Mezei <jfmezei.spamnot@videotron.ca> >wrote:e >d >>Jim Tauberg wrote:G >>> installed MOTIF and after the install was asked to reboot.  Now the F >>> system comes up in a graphical format, but when I try to log in asF >>> SYSTEM, it shows a dialog that says "No license is active for this >>> software product"n >>? >>Press the reset button or break key to get to the >>> prompt.- >> >>Then B/1 to get to SYSBOOT >>! >>at SYSBOOT> SET WINDOW_SYSTEM 0s >>    SYSBOOT> CONTINUEo >>L >>This will bring up your machine without any windowing and just a very veryN >>basic character cell interface (not even VT100, more like TTY). You can thenG >>type in the licence information, then use SYSGEN to USE CURRENT , SET 2 >>WINDOW_SYSTEM 1 , WRITE CURRENT and then reboot. >DJ >Decwindows has a habit of just starting anyway, ime.  (If you look at theM >startup procedure, you'll see it assumes that WINDOW_SYSTEM set to zero must  >mean 1 of course.)d > M >Probably easier to do a minimum boot, although this may get you into troubleo< >if you rely too much on logical names that are not defined: >e >SYSBOOT> SET STARTUP_P1 "MIN" >SYSBOOT> SET WRITESYSPARAMS 0 >SYSBOOT> CONT > J >The second line will prevent an implicit WRITE CURRENT, so you can simply >reboot normally afterwards. >s >y >	John   ------------------------------  % Date: Fri, 19 Apr 2002 12:50:34 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>rA Subject: Re: Stuck in licenseless Motif with OpenVMS 7.2 Hobbyisti, Message-ID: <3CC04AD8.88632A7D@videotron.ca>   Jason O'Donnell wrote:D > You may also get layered product licenses at the same site you gotH > your hobbyist license.  It includes motif licenses.  Heck you can even > get ALL-IN-1. ;)  N Nop. ALL-IN-1 isn't included with the hobbyist layeyed products. Only the veryN old/retired "all-in-1 mail" stuff which is not related to ALL-IN-1 in any way.   ------------------------------  # Date: Fri, 19 Apr 2002 16:02:42 GMTr# From: "John Smith" <a@nonymous.com> 9 Subject: Re: Sue, a great promotion for "unhackable" vms!oF Message-ID: <CcXv8.9574$Gpn.3655@news02.bloor.is.net.cable.rogers.com>  9 "Paul Repacholi" <prep@prep.synonet.com> wrote in messagee' news:87hem8y9pp.fsf@prep.synonet.com...:H > winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr") writes:o >cG > > They'd do better working up something with a quirky band like _TheysE > > Might Be Giants_, who'd probably be amused to write a VMS jingle. G > > (Actually, to reach CEOs rather than young hackers, they should gettG > > Elton John to rewrite - no, not Candle in the Wind again - but "I'm D > > still standing", and make a video showing huge disasters and VMSH > > clusters staying up.  They could point out that John himself was bigC > > in the 70s and is still big, like VMS.)  A lot of the people on,D > > corporate boards listened to him when they were younger and they! > > still vaguely think he's hip.b >sF > I was thinking of two Alphas standing after the dust cloud clears...F > You would only need to run it once, the yelling and screaming should. > keep it going for at least another 6 months. >   G Like the Macintosh 1984 'Big Brother' Super bowl commercial. People areo still talking about that.   4 But one has to keep in mind that there's sometimes aJ difference in advertising that wins Clio awards and advertising that sells product.   ------------------------------  % Date: Fri, 19 Apr 2002 10:32:03 +0100I4 From: John Laird <john@laird-towers.freeserve.co.uk>' Subject: Re: System Disk as Quorum Disk 8 Message-ID: <k6mvbu4dvun6htgplkhvv8o1o7vs02kqhr@4ax.com>  H On Thu, 18 Apr 2002 10:34:36 +0100, "Craig Cooke" <ccooke@beta.dabs.com> wrote:  G >I've been away from VMS (oops sorry OpenVMS) for a few years, back now,M >though.  Started working for a e-tailor recently, had a look at their 2 node E >system which they thought was clustered - well, the sysman parameter K >VAXCLUSTER is set to 2. - But then EXPECTED_VOTES is set to 1 and there is. >no Quorum Disk. M  G That won't stop a cluster forming.  Equally, it won't stop two clusters J forming if they can't see one another over a suitable network interconnectL when they both boot up.  Or if one crashes, then the other will hang, having lost quorum.  I Pick a disk you couldn't really live without for the quorum disk.  With 2oK nodes, you're hardly going to be impacted by the extra I/O (it's about oncee per second or so, IIRC).     	Johnr -- e
 John Laird Yezerski Roper Ltd   ------------------------------  # Date: Fri, 19 Apr 2002 15:35:36 GMT-7 From: tauberg@nauticom.no_spam_please.net (Jim Tauberg) + Subject: TCP/IP 5.0 Configuration QuestionsR1 Message-ID: <3cc13732.53581228@news.nauticom.net>D   Hi,8> 	I've installed OpenVMS 7.2 from the Hobbyist CD and have alsoD installed MOTIF and am now working on the TCP/IP 5.0 from the cd.  IA installed it and ran the configuration utility and configured thevE interface under Core Environment.  I've been able to Ping myself (hey - baby) and the other machines on this network.-  B I'm trying to configure the system to reach the internet through a= gateway (dsl router) on my network and also to set up the DNSp
 entries...  < I took a look through the manual and have tried things like:    8 SET NAME_SERVICE /SERVER=xxx.xxx.xxx.xxx /ENABLE /SYSTEM  - but this returns:  REQUIRES TCP-IP-CLIENT PAK)  - What is the client pak and where do I get it?,  ? What commands will I need to set my default gateway and my dns?e   Thanks in advance, Jim Tauberg  tauberg@no_spam.nauticom.net   ------------------------------   Date: 19 Apr 02 08:17:07 +0200) From: p_sture@elias.decus.ch (Paul Sture)o5 Subject: Re: Time to finally rename VMS back to VMS ?W) Message-ID: <zqv6rhc67Vq4@elias.decus.ch>   c In article <3CBFADDB.3DB0A03B@aaa.com>, Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> writes:  > The link gives me :i( > "550 file specification syntax error". >  > Jan-Erik Sderholm.d >  > Paul Sture wrote:  >> -, >> Talking of things leaping out of screens: >> 0# >> ftp://80.254.165.138/mil_bug.gife >> (animated GIF - 77KB) >>  6 Sorry. It worked for me. It could be an ODS-5 problem. Now changed to ODS-2.P  ! What were you using to access it?s   __
 Paul Sture Switzerland   ) Finally, I can truthfully steal this sig:t  5 The box said Windows 98 or better, so I bought a Mac.n   ------------------------------  % Date: Fri, 19 Apr 2002 08:24:34 +0200P9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>i5 Subject: Re: Time to finally rename VMS back to VMS ? ' Message-ID: <3CBFB822.784F9027@aaa.com>s  # Yes, that did a change. Now I get :g  "550 File not found" instead :-)  % I'm using Netscape Communicator 4.74 e  	 Jan-Erik.I   Paul Sture wrote:r > e > In article <3CBFADDB.3DB0A03B@aaa.com>, Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> writes:] > > The link gives me : * > > "550 file specification syntax error". > >I8 > Sorry. It worked for me. It could be an ODS-5 problem. > Now changed to ODS-2.P > # > What were you using to access it?P >1   ------------------------------  % Date: Fri, 19 Apr 2002 12:40:19 -04001- From: JF Mezei <jfmezei.spamnot@videotron.ca> 5 Subject: Re: Time to finally rename VMS back to VMS ?P, Message-ID: <3CC04872.E891AF18@videotron.ca>   Paul Sture wrote:$% > >> ftp://80.254.165.138/mil_bug.gifY > >> (animated GIF - 77KB)  9 When I tried the above yesterday on netscape,  it failed.1   However, I was able to:I    ftp://80.254.165.138   J which displayed that one file and then clicking on the file, I was able to view it.   ------------------------------  " Date: Fri, 19 Apr 02 11:19:21 +100 From: rok@nuk.uni-lj.sis$ Subject: Two DMQ COM servers running& Message-ID: <3cbfef2f.0@NUK.Uni-Lj.Si> Keywords: DMQ, MCS  I AlphaServer 4100 5/400 4MB, OpenVMS AXP V7.3, DMQ V3.2A, Process Software  MultiNet V4.2 Rev A-XS  $  From DMQ$COM_SERVER_0202_00047.LOG:2 DmQ T 00:56.3 Time Stamp - 19-APR-2002 10:00:56.324 DmQ T 00:56.3 ------- COM Server Starting ----------J DmQ I 00:56.3 COM Server (V3.2A-22(3222)) starting at 19-APR-2002 10:00:56? DmQ I 00:56.3 MCS already exists - two DMQ COM servers running.m  D  In contrast to this statement SHOW SYSTEM shows *no* DMQ COM server is running.lA  Reboot does clear DMQ's false notion; does anybody know a milder 	 solution?_   Regards,  D Rok Vidmar                       Internet:  rok.vidmar@nuk.uni-lj.si; National and University Library  Phone:     +386 1 421 5461c; Turjaska 1, SI-1000 Ljubljana    Fax:       +386 1 421 5464d Slovenia   ------------------------------  % Date: Fri, 19 Apr 2002 20:49:04 +0800y- From: David B Sneddon <dbsneddon@bigpond.com>_( Subject: Re: Two DMQ COM servers running( Message-ID: <02041920490400.02149@bwian>  3 On Thu,  1 Jan 1970 08:00, rok@nuk.uni-lj.si wrote:rK > AlphaServer 4100 5/400 4MB, OpenVMS AXP V7.3, DMQ V3.2A, Process Software1 > MultiNet V4.2 Rev A-X. >]& >  From DMQ$COM_SERVER_0202_00047.LOG:4 > DmQ T 00:56.3 Time Stamp - 19-APR-2002 10:00:56.326 > DmQ T 00:56.3 ------- COM Server Starting ----------L > DmQ I 00:56.3 COM Server (V3.2A-22(3222)) starting at 19-APR-2002 10:00:56A > DmQ I 00:56.3 MCS already exists - two DMQ COM servers running.  >mF >  In contrast to this statement SHOW SYSTEM shows *no* DMQ COM server
 > is running.yC >  Reboot does clear DMQ's false notion; does anybody know a milder6 > solution?n >d
 > Regards, >lF > Rok Vidmar                       Internet:  rok.vidmar@nuk.uni-lj.si= > National and University Library  Phone:     +386 1 421 54610= > Turjaska 1, SI-1000 Ljubljana    Fax:       +386 1 421 5464-
 > Slovenia   Rok,  F DmQ uses locks to control a lot of its activities.  What you will findE is that there is likely a "hung" process somewhere on the system that]B was connected to DmQ but for some reason is in a state where it isF unable to release the DmQ locks.  You will have to locate this process  and make it release the locks...G The DmQ locks all begin with DMQ.  We sometimes have this situation butcH I have a program to help out.  I do "$ getlki dmq*" and it shows me whatD processes have the DmQ locks.  SDA can also be used but I have neverE used it to look at locks.  Send me e-mail if you would like a copy ofA GETLKI.    Regards, Dave.f -- II David B Sneddon (dbs)  OpenVMS Systems Programmer   dbsneddon@bigpond.com_I Sneddo's quick guide ...          http://www.users.bigpond.com/dbsneddon/dI DBS freeware at ...   http://www.users.bigpond.com/dbsneddon/software.htmnI "Life is what happens to you while you're busy making other plans" Lennon0   ------------------------------  % Date: Fri, 19 Apr 2002 15:40:49 -0000 - From: wspencer@ap.nospam.org (Warren Spencer)f( Subject: Re: Two DMQ COM servers running7 Message-ID: <91F57F058warrenspencer1977@209.249.90.100>.  6 rok@nuk.uni-lj.si wrote in <3cbfef2f.0@NUK.Uni-Lj.Si>:  J >AlphaServer 4100 5/400 4MB, OpenVMS AXP V7.3, DMQ V3.2A, Process Software >MultiNet V4.2 Rev A-X >S% > From DMQ$COM_SERVER_0202_00047.LOG:S3 >DmQ T 00:56.3 Time Stamp - 19-APR-2002 10:00:56.32 5 >DmQ T 00:56.3 ------- COM Server Starting ---------- K >DmQ I 00:56.3 COM Server (V3.2A-22(3222)) starting at 19-APR-2002 10:00:56e@ >DmQ I 00:56.3 MCS already exists - two DMQ COM servers running. >0E > In contrast to this statement SHOW SYSTEM shows *no* DMQ COM server; >is running.B > Reboot does clear DMQ's false notion; does anybody know a milder
 >solution? >0	 >Regards,  > E >Rok Vidmar                       Internet:  rok.vidmar@nuk.uni-lj.siS< >National and University Library  Phone:     +386 1 421 5461< >Turjaska 1, SI-1000 Ljubljana    Fax:       +386 1 421 5464	 >Slovenia.  J That message is a bit of a fib - it's no necessarily another COM server - L it may also be a user process that uses DMQ.  Do you know on which disk the I dmq group's files reside?  Usually a $ SHOW DEV/FILES Dxxx will show you eK which process has DMQ files mapped.  Stop that process (or processes), and e+ you'll be able to restart dmq successfully.-   ws   -- ,   Warren Spencer' Senior Software Engineer (not a writer)o The Associated Press  < ** Time flies like an arrow.  Fruit flies like a bananna. **   ------------------------------  % Date: Fri, 19 Apr 2002 01:11:14 -0400n( From: David Froble <davef@tsoft-inc.com>5 Subject: Re: UNXSIGNAL bugcheck on virtual disk, why?a, Message-ID: <3CBFA6F2.6020705@tsoft-inc.com>   Lawrence Bleau wrote:e  N > Hi, I just has a system crash, bugcheck UNXSIGNAL, Unexpected signal name inO > ACP.  I'm running OpenVMS AXP 7.1-2.  I just added a 180Gb disk to my system,eQ > and was using Glenn Everhart's virtual disk driver, VDDRIVER, V4f, to partitione > it.  > I > Let me back up and give you the sequence.  The disk went in without anyrM > problems, showed up on SHO DEV properly.  I did an INIT first, which workedrO > fine.  I then did a MOUNT/FOREIGN; again, no problems.  I connected a virtualoQ > disk device (VDA20:) and assigned it a block range starting at LBN 0 and having-. > size 97693755 (50Gb); the command I used was > 4 > $ asnvdm6/assign/lbn=0/len=97693755 vda20: dkc500: > C > I then did an INIT on the VD device, and got the following error:H > Q > %INIT-F-INTDIV, arithmetic trap, integer divide by zero at PC=FFFFFFFF80007A18,0
 > PS=0000000B0 > O > After puzzling over this a bit, I recalled that Glenn's recommendation was tolQ > use /NOVEREIFY, so I did an INIT/NOVERIFY VDA20: BLANK and it worked.  So, I've Q > dispatched the initial problem (I hope; maybe not, that's why I'm telling you).  > A SHOW DEV VDA20 gives > O > Disk SAMPEX$VDA20:, device type Foreign disk type 1, is online, file-orientedt >     device, shareable. > Q >     Error count                    0    Operations completed                  03Q >     Owner process                 ""    Owner UIC                      [SYSTEM]sQ >     Owner process ID        00000000    Dev Prot            S:RWPL,O:RWPL,G:R,WnQ >     Reference count                0    Default buffer size                 512WQ >     Total blocks            97693696    Sectors per track                    32mQ >     Total cylinders            29868    Tracks per cylinder                  32E > N > I did a normal mount on the initialized VD device, which succeeded.  Next, IN > did the command CREATE/DIRECTORY VDA20:[LDF] to create a top-level directoryQ > for my data.  *This* is what caused the system to crash!  Here is selected info  > from the crash dump: > = > Bugcheck Type:     UNXSIGNAL, Unexpected signal name in ACPu > VMS Version:       V7.1-2  > Current Process:   SYSTEM-D > Current Image:     SAMPEX$DKA0:[SYS0.SYSCOMMON.][SYSEXE]CREATE.EXE< > Failing PC:        FFFFFFFF.80007A18    OTS$REM_UL_C+000B8& > Failing PS:        10000000.00000000# > Module:            SYS$BASE_IMAGEe > Offset:            00005A18r > N > I can provide other info as needed.  Alternately, I've made the listing fileK > VMS generates upon booting after a crash available: go to sampex.umd.edu,s- > anonymous ftp, and get the file crash.lis .y > O > I should hasten to add that I've been using VDDRIVER on this system for yearssP > now without problems.  I did not change VDDRIVER version or VMS version; all IM > did is add a bigger disk, reboot, and try to use it.  Btw, a CRE/DIR on thehQ > underlying disk, when mounted as a normal Files-11, works, as does a CRE/DIR onpJ > a 50Gb physical disk I already have (almost exact same block size, too).	 > Thanks.e >  > Lawrence Bleau > University of Maryland$ > Physics Dept., Space Physics Group > 301-405-6223 > bleau@umtof.umd.edu  >   K Some idle speculation.  Keep in mind that the largest disk that can have a  L clustersize of 1 in somewhere around 137 or 138 GB.  I'm guessing that this Q limit may be based upon the limitations of the LONGWORD integer.  Doing the math e6 in my head caused a headache, so I quit that exercise.  M Possibly the large drive is causing a longword overflow in VDDRIVER?  Just a c) guess, and no real effort to do the math.c   Dave   ------------------------------  % Date: Fri, 19 Apr 2002 09:51:06 -0400.1 From: "Mark D. Jilson" <jilly@clarityconnect.com>h5 Subject: Re: UNXSIGNAL bugcheck on virtual disk, why? 2 Message-ID: <3CC020CA.F2980303@clarityconnect.com>  H This crash has been seen before with large disks and has come down to anD overflow of UCB$W_CYCLINDERS.  Patch information is needed to see if& there might be a way to get past this.   David Froble wrote:e >  > Lawrence Bleau wrote:y > P > > Hi, I just has a system crash, bugcheck UNXSIGNAL, Unexpected signal name inQ > > ACP.  I'm running OpenVMS AXP 7.1-2.  I just added a 180Gb disk to my system,rS > > and was using Glenn Everhart's virtual disk driver, VDDRIVER, V4f, to partitiono > > it.  > >>K > > Let me back up and give you the sequence.  The disk went in without anyrO > > problems, showed up on SHO DEV properly.  I did an INIT first, which worked Q > > fine.  I then did a MOUNT/FOREIGN; again, no problems.  I connected a virtual S > > disk device (VDA20:) and assigned it a block range starting at LBN 0 and having 0 > > size 97693755 (50Gb); the command I used was > >t6 > > $ asnvdm6/assign/lbn=0/len=97693755 vda20: dkc500: > > E > > I then did an INIT on the VD device, and got the following error:c > >.S > > %INIT-F-INTDIV, arithmetic trap, integer divide by zero at PC=FFFFFFFF80007A18,o > > PS=0000000Bo > > Q > > After puzzling over this a bit, I recalled that Glenn's recommendation was toTS > > use /NOVEREIFY, so I did an INIT/NOVERIFY VDA20: BLANK and it worked.  So, I've S > > dispatched the initial problem (I hope; maybe not, that's why I'm telling you).- > > A SHOW DEV VDA20 gives > >eQ > > Disk SAMPEX$VDA20:, device type Foreign disk type 1, is online, file-oriented  > >     device, shareable. > > S > >     Error count                    0    Operations completed                  0oS > >     Owner process                 ""    Owner UIC                      [SYSTEM]gS > >     Owner process ID        00000000    Dev Prot            S:RWPL,O:RWPL,G:R,WLS > >     Reference count                0    Default buffer size                 512hS > >     Total blocks            97693696    Sectors per track                    32.S > >     Total cylinders            29868    Tracks per cylinder                  32h > >tP > > I did a normal mount on the initialized VD device, which succeeded.  Next, IP > > did the command CREATE/DIRECTORY VDA20:[LDF] to create a top-level directoryS > > for my data.  *This* is what caused the system to crash!  Here is selected infoc > > from the crash dump: > >N? > > Bugcheck Type:     UNXSIGNAL, Unexpected signal name in ACP  > > VMS Version:       V7.1-2o > > Current Process:   SYSTEM F > > Current Image:     SAMPEX$DKA0:[SYS0.SYSCOMMON.][SYSEXE]CREATE.EXE> > > Failing PC:        FFFFFFFF.80007A18    OTS$REM_UL_C+000B8( > > Failing PS:        10000000.00000000% > > Module:            SYS$BASE_IMAGEb > > Offset:            00005A18d > > P > > I can provide other info as needed.  Alternately, I've made the listing fileM > > VMS generates upon booting after a crash available: go to sampex.umd.edu,l/ > > anonymous ftp, and get the file crash.lis .e > >hQ > > I should hasten to add that I've been using VDDRIVER on this system for years R > > now without problems.  I did not change VDDRIVER version or VMS version; all IO > > did is add a bigger disk, reboot, and try to use it.  Btw, a CRE/DIR on theaS > > underlying disk, when mounted as a normal Files-11, works, as does a CRE/DIR onPL > > a 50Gb physical disk I already have (almost exact same block size, too). > > Thanks.s > >i > > Lawrence Bleau > > University of Maryland& > > Physics Dept., Space Physics Group > > 301-405-6223 > > bleau@umtof.umd.edu1 > >  > L > Some idle speculation.  Keep in mind that the largest disk that can have aM > clustersize of 1 in somewhere around 137 or 138 GB.  I'm guessing that thiskR > limit may be based upon the limitations of the LONGWORD integer.  Doing the math8 > in my head caused a headache, so I quit that exercise. > N > Possibly the large drive is causing a longword overflow in VDDRIVER?  Just a+ > guess, and no real effort to do the math.y >  > Dave   -- rC Jilly	- Working from Home in the Chemung River Valley - Waverly, NY 0 	- jilly@clarityconnect.com			- Brett Bodine fan. 	- Mark.Jilson@Compaq.com			- since 1975 or so, 	- http://www.jilly.baka.com               -   ------------------------------    Date: 19 Apr 2002 01:48:51 -07003 From: Eric Smith <eric-no-spam-for-me@brouhaha.com>o Subject: Re: VAX/Alpha CIt0 Message-ID: <qhn0w0dqfg.fsf@ruckus.brouhaha.com>  . Paul Repacholi <prep@prep.synonet.com> writes:E > CI cables are identical to 10Base5, but with TNC connectors insteade > of N.o  ) So it's simply RG-58 with TNC connectors?s   ------------------------------    Date: 19 Apr 2002 07:06:22 -0700- From: jodonnell@hrblock.com (Jason O'Donnell)m Subject: Re: VAX/Alpha CI = Message-ID: <9059bf6b.0204190606.31e6bee2@posting.google.com>3  a Paul Repacholi <prep@prep.synonet.com> wrote in message news:<874ri9ylts.fsf@prep.synonet.com>...l1 > jodonnell@hrblock.com (Jason O'Donnell) writes:P > G > > Also, all cables were cut to the coupler, so I will need to get theiA > > CI cables and rewire the power input cable.  Any suggestions?n > E > CI cables are identical to 10Base5, but with TNC connectors insteadgD > of N. The is no power connection to the SC. BTW, do not let the SC4 > body ground out. Keep it isolated from everything. > . > Cutting the cables... What sort of idiots...  * Thanks, this is really useful information.  : They were decommissioning the equipment and didn't care...  E When I opened up the SC last night I noticed two difference thickness,C of cables being used.  They seemed to be paired in the connections.   C Also, the storageworks cabinet with the dual HSJ40 controllers does-= not appear to have the same CI connector as the SC.  The onlyrD connector I see is on the front of each controller and it looks like this:    ---------------- \  o o    o o  /  --------------u  $ Am I not looking in the right place?   ------------------------------   Date: 19 Apr 02 10:49:51 EDT From: grant@rigel.cc.wmich.edu Subject: Re: VAX/Alpha CIt/ Message-ID: <nzpymmQnnDmS@mcduck.acs.wmich.edu>l  m In article <9059bf6b.0204190606.31e6bee2@posting.google.com>, jodonnell@hrblock.com (Jason O'Donnell) writes:>c > Paul Repacholi <prep@prep.synonet.com> wrote in message news:<874ri9ylts.fsf@prep.synonet.com>...O2 >> jodonnell@hrblock.com (Jason O'Donnell) writes: >> SH >> > Also, all cables were cut to the coupler, so I will need to get theB >> > CI cables and rewire the power input cable.  Any suggestions? >> -F >> CI cables are identical to 10Base5, but with TNC connectors insteadE >> of N. The is no power connection to the SC. BTW, do not let the SCn5 >> body ground out. Keep it isolated from everything.a >> m/ >> Cutting the cables... What sort of idiots...F > , > Thanks, this is really useful information. > < > They were decommissioning the equipment and didn't care... > G > When I opened up the SC last night I noticed two difference thicknessaE > of cables being used.  They seemed to be paired in the connections.o > E > Also, the storageworks cabinet with the dual HSJ40 controllers doesb? > not appear to have the same CI connector as the SC.  The onlysF > connector I see is on the front of each controller and it looks like > this:0 >  > ---------------- > \  o o    o o  / >  --------------R > & > Am I not looking in the right place?  G You are looking in the right place. The cable that plugs into that plugfH on the front of the controller drops under the floor and connects to theJ four CI cables.  The eliminates the necessity of getting four thick, stiffI cables (or thin, potentially fragile cables) threaded up into the cabinetnJ and through a VERY TIGHT (i.e., will break the cable) bend to come off the front of the controller.  E I have a few lying around since we converted to fiber. (also one pair F of HSJ40s). I can look up the cable number if you need it.  PresumablyF the people who deinstalled the system have the cables -- or threw them out or destroyed them.   ------------------------------    Date: 19 Apr 2002 10:11:16 -0700- From: jodonnell@hrblock.com (Jason O'Donnell)r Subject: Re: VAX/Alpha CI = Message-ID: <9059bf6b.0204190911.4388cd09@posting.google.com>0  I > You are looking in the right place. The cable that plugs into that plug J > on the front of the controller drops under the floor and connects to theL > four CI cables.  The eliminates the necessity of getting four thick, stiffK > cables (or thin, potentially fragile cables) threaded up into the cabinetkL > and through a VERY TIGHT (i.e., will break the cable) bend to come off the > front of the controller. > G > I have a few lying around since we converted to fiber. (also one pairaH > of HSJ40s). I can look up the cable number if you need it.  PresumablyH > the people who deinstalled the system have the cables -- or threw them > out or destroyed them.  F If you can get me the part numbers, that would be wonderful.  Also, ifD you can part with the cables, I would pay for them.  You're in K-ZooC right?  How about the HSJ40s?  It appears that you can only run six.F 350 racks for each pair.  I have ten racks.  I still have read more of the documentation.  ! Anyone have SC and/or HSJ40 docs?-   ------------------------------  % Date: Fri, 19 Apr 2002 09:15:48 +0200s2 From: "Ren Schelbaum" <rene.schelbaum@datakom.at>0 Subject: Re: VMS Cluster address switching issueG Message-ID: <3cbfc410$0$19168$5039e797@newsreader01.highway.telekom.at>n  ; "Rainer Lehrig" <lehrig@t-online.de> schrieb im Newsbeitragi% news:3CBFA077.D7EDF4F9@t-online.de...o > Hi,r >k? > working as a freelancer I'm faced with the following problem.4 >5H > There is a small cluster with 2 machines (Alpha, OpenVMS 7.2-1, a RAIDA > connected to each machine, an ethernet adapter on each machine).; > We want 1 machine to be the master and 1 to be the slave.sC > When the slave becomes master all network communication has to bes' > switched to the former slave machine.eH > For testing we switch the cluster address with "UCX SET INTERFACE ..."  > from one machine to the other. >  > Observation:I > Our network applications reconnect within 5 seconds after we change the  > cluster address. > This is ok for us.G > But sometimes we have to wait 30-40 seconds before messages arrive at : > the destination although the applications are connected.- > In some cases all is done within 5 seconds. C > After this time expires, we get all the messages from these 30-40n > seconds at once.) > After this the communication is normal.s >d > Question:e8 > Has anybody an idea about the delay of 30-40 seconds ? >. > Your:f > Rainer Lehrig    Hi!w  ! Maybe ARP is getting in your way.LG The other machines connecting to your server (or routers via which they4I connect) have the hardware address of your ethernet-port (MAC-address) inDI their ARP-cache. As the MAC-Address changes with the switchover, it takes1F some time until this information times out and is being refreshed. TheL fastest way to solve this problem is by sending a ping to all servers and/or routers via which they connect.M   Ren   ------------------------------  % Date: Fri, 19 Apr 2002 07:59:15 -0400Q5 From: David Beatty <David.Beatty@qwertysasasdfgh.com>Q0 Subject: Re: VMS Cluster address switching issue2 Message-ID: <RwbAPJh2xWgSdyqfChb8=Sant09+@4ax.com>  C     Just out of curiousity, is there any reason you are not runningn8 the cluster alias under Compaq TCP/IP Services V5.0?  If< my memory serves me correctly, V4.2 is not supported on V7.2 and higher of VMS.   David R. Beattyk  F On Fri, 19 Apr 2002 06:43:35 +0200, Rainer Lehrig <lehrig@t-online.de> wrote:   >Hi, >L> >working as a freelancer I'm faced with the following problem. > G >There is a small cluster with 2 machines (Alpha, OpenVMS 7.2-1, a RAIDb@ >connected to each machine, an ethernet adapter on each machine): >We want 1 machine to be the master and 1 to be the slave.B >When the slave becomes master all network communication has to be& >switched to the former slave machine.G >For testing we switch the cluster address with "UCX SET INTERFACE ..."  >from one machine to the other.a >n
 >Observation:rH >Our network applications reconnect within 5 seconds after we change the >cluster address.i >This is ok for us.aF >But sometimes we have to wait 30-40 seconds before messages arrive at9 >the destination although the applications are connected.:, >In some cases all is done within 5 seconds.B >After this time expires, we get all the messages from these 30-40 >seconds at once.t( >After this the communication is normal. >p
 >Question:7 >Has anybody an idea about the delay of 30-40 seconds ?a >d >Your: >Rainer Lehrig   ------------------------------  % Date: Fri, 19 Apr 2002 09:46:06 -0700t% From: Dean Woodward <deanw@rdrop.com> & Subject: [Fwd: New OpenVMS Whitepaper]) Message-ID: <3CC049CE.CF5317A9@rdrop.com>    "Shadrake, Tony" wrote:  > C > We have just published a new technical case study on extending an G > OpenVMS client/server application to the Internet and I would like to C > send you a copy.  This case study is based on Compaq BASEstar and) > Attunity Connect.w  F Anybody else get spam from this, this, this.... Sorry, mum always told' me "if you can't say something nice..."e  B If anyone out there knows Mr. Shadrake, send him a clue- and Sue's address.   ------------------------------   End of INFO-VAX 2002.216 ************************