0 INFO-VAX	Mon, 09 Jan 2006	Volume 2006 : Issue 18      Contents: Apache 2.1 running on OVMS IA64 ' Re: Beginner: nesting DCL command files  DS10 systems discounts !/ Re: foreign commands, .com files, and dcltables / Re: foreign commands, .com files, and dcltables / Re: foreign commands, .com files, and dcltables / Re: foreign commands, .com files, and dcltables & Re: Free to a good home: OSF/1 manuals& Re: Free to a good home: OSF/1 manuals# Re: Java 1.4.2 installation problem  Re: KVM Switches Re: more lbr$flush problems  Re: more lbr$flush problems  Re: more lbr$flush problems  Re: more lbr$flush problems  Re: more lbr$flush problems  Re: more lbr$flush problems ) Re: Need model number for Infoserver-1000 ! Re: old EZ disks with latest VMS? ! Re: old EZ disks with latest VMS? - Recommendations for 4mm tape on 4000/600 Vax? 1 Re: Recommendations for 4mm tape on 4000/600 Vax? ! Re: Round internal SCSI cables ??   Re: UNIX shm* functions for VMS?  Re: UNIX shm* functions for VMS?' Re: Using VT420 as Alpha System Console ' Re: Using VT420 as Alpha System Console ' Re: Using VT420 as Alpha System Console + Re: WVNETcluster uptime reaches 10 years... + Re: WVNETcluster uptime reaches 10 years...   F ----------------------------------------------------------------------  $ Date: Mon, 9 Jan 2006 15:04:35 +00005 From: sysulfo <sysulfo.21djh0@nomx.sysadminforum.com> ( Subject: Apache 2.1 running on OVMS IA643 Message-ID: <sysulfo.21djh0@nomx.sysadminforum.com>    Hi !  A Does anyone have any experience regarding running CSWS/Apache 2.1 + on Integrity, OpenVMS 8.2-1 AND Rdb 7.2 ???   1 We are in great need of help to get this working. 
 Config ...
 Client(s) ...  PHP ...    Rgds.    Ulf Ohlsson      --   sysulfo " posted via http://sysdminforum.com   ------------------------------  % Date: Mon, 09 Jan 2006 11:13:14 +0200 4 From: Mike Rechtman <michael.rechtman.nospam@hp.com>0 Subject: Re: Beginner: nesting DCL command files& Message-ID: <43C2454A.47E9BE49@hp.com>   Alder wrote: > I > I've added command files to my SYSTARTUP_VMS.COM file to define various B > logicals and symbols and to call other command files that defineD > logicals and symbols.  I know the first level of command files getD > called, but the second-level command files are not.  Is this to beF > expected for some reason?  This is OVMS 7.3-2 and I did RTFM.  Did I. > miss something about nesting DCL procedures? > I > One of the second-level files is the unedited TCPIP$DEFINE_COMMANDS.COM H > shipped with VMS, so I imagine its coding is not the problem.  AnotherG > of these second-level command files is the one shipped with SAMBA for I > VMS.  After startup the logicals and symbols (foreign commands) defined B > in these two second-level files remain undefined.  Below are the# > relevant snippets from the files.  > H > Can anyone help with a pointer on where I went wrong, or what to read?	 > Thanks.  >  > Terry  >  > SYSTARTUP_VMS.COM  > =================  > ... A > $ DEFINE/SYS/EXEC/TRANS=(CONCEALED)SAMBA_ROOT DISK$NET:[SAMBA.]  > ... ! > $ @SYS$STARTUP:STARTUP_PRODUCTS  > , > SYS$STARTUP:STARTUP_PRODUCTS.COM (level 1) > ================== > ... & > $ @SYS$STARTUP:TCPIP$DEFINE_COMMANDS# > $ @SAMBA_ROOT:[BIN]SAMBA_COMMANDS  > ...  >  > SAMBA_COMMANDS.COM (level 2) > ================== > ... ) > $procedure = f$environment("PROCEDURE")  > $! > $command = "CONVERT_TDB" > $program = "''command'.EXE;0" . > $'command' :== $'f$parse(program,procedure)' > ...   - I'm not familiar with SAMBA_COMMANDS.COM, but F SYS$STARTUP:TCPIP$DEFINE_COMMANDS is supposed to be run (activated) byA *every user* that wants to use TCPIP commands (ping, nslookup...) , If its a small number of users, add the line%  $ @SYS$STARTUP:TCPIP$DEFINE_COMMANDS 7 to each LOGIN.COM. If its for everyone, add the line to 5 SYS$MANAGER:SYLOGIN.COM in the "INTERACTIVE" section.    Mike --  E --------------------------------------------------------------------- E Usual disclaimer: All opinions are mine alone, perhaps not even that. ? Mike Rechtman                            *rechtman@tzora.co.il* F Kibbutz Tzor'a.                          Voice (home): 972-2-9908337  B   "20% of a job takes 80% of the time, the rest takes another 80%"E ---------------------------------------------------------------------  -----BEGIN GEEK CODE BLOCK-----  Version: 3.1: GCM/CS d(-)pu s:+>:- a++ C++ U-- L-- W++ N++ K? w--- V+++$6 PS+ PE-- t 5? X- tv-- b+ DI+ D-- G e++ h--- r+++ y+++@ ------END GEEK CODE BLOCK------    ------------------------------  $ Date: Mon, 9 Jan 2006 10:42:34 -0500C From: "David Turner, Island Computers US Corp" <dbturner@icusc.com> ! Subject: DS10 systems discounts ! : Message-ID: <Gnvwf.27368$0y2.10393@bignews2.bellsouth.net>  ! Alphaserver DS10 systems in stock      Alphaserver DS10 466Mhz  CDROM and Floppy 512MB Memory 18GB U160 Disk! U160 Dual Channel SCSI Controller  Dual Ethernet On Board   Without licenses $2995   With OpenVMS Base License $3700 ( With OpenVMS Base and EIP Licenses $4600     Alphaserver DS10 617Mhz  CDROM and Floppy 512MB Memory 18GB U160 Disk! U160 Dual Channel SCSI Controller  Dual Ethernet On Board   Without licenses $3995   With OpenVMS Base License $4700 ( With OpenVMS Base and EIP Licenses $5600     While stocks last   - Mention COMP.OS.VMS Special of January 9th 06       < Everything comes with a 12 month Island Computers warranty !     David Turner Island Computers US Corp 2700 Gregory St, Suite 180 Savannah GA 31404  Tel: 912 447 6622 X201 Cell: 912 447 6622 X252  Fax: 912 201 0402  Email: dbturner@icusc.com  Web: http://www.islandco.com% ===================================== < All orders are subject to the following terms and conditions. of sale. These should be read before ordering.% http://www.islandco.com/warranty.html    ------------------------------  % Date: Mon, 09 Jan 2006 04:04:40 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 8 Subject: Re: foreign commands, .com files, and dcltables, Message-ID: <43C226EC.4E4A250E@teksavvy.com>   mabbuttg@yahoo.ca wrote:6 > our users' login scripts, similar to 'NEWPROGRAM :== > "@[DIR]NEW_PROGRAM" '.  I > What I've found is some info on modifying the global DCLTABLES, using a ' > .cld file, with the sample syntax of:   G You cannot add a command procedure to the DCLTABLES. This works only on F images. (.EXE files), and when defining a CLD for it, the image itself< needs to call special VMS routines to extract the arguments.  C What you can do however is automatically add the definition of this B process symbol in SYS$MANAGER:SYLOGIN.COM as this file is executedF before the user's LOGIN.COM and this is where you can make definitions that will apply to all users.   F You could further refine this by acting on the user's group number andG from SYLOGIN.COM, calling a group specific command procedure that would ( make definitions specific to that group.  G > - if for some reason I wanted to remove this entry from the DCLTABLES & > after adding it, how can I do that?   < SET COMMAND/DELETE=mycommand/TABLE=SYS$LIBRARY:DCLTABLES.EXE   ------------------------------  % Date: Mon, 09 Jan 2006 11:06:08 +0200 4 From: Mike Rechtman <michael.rechtman.nospam@hp.com>8 Subject: Re: foreign commands, .com files, and dcltables& Message-ID: <43C243A0.4E1DBC50@hp.com>   mabbuttg@yahoo.ca wrote: > - > I'm fairly new to OpenVMS, so bear with me:  > F > A consultant implemented a process for us where several programs nowC > call a .com file with arguments to complete some file operations.  > I > To make this work, we had to add a foreign command symbol definition to 6 > our users' login scripts, similar to 'NEWPROGRAM :== > "@[DIR]NEW_PROGRAM" '. > E > While we've gone through and added this symbol to what we're pretty C > sure are all the users who are affected by this, I was doing some H > research to see what could be done on a global level just to make sureD > (in our setup it doesn't matter if all users have the above symbolI > defined, as we have additional access restrictions on the programs that 4 > call it, and most users don't have prompt access). > I > What I've found is some info on modifying the global DCLTABLES, using a ' > .cld file, with the sample syntax of:  >  > DEFINE VERB FOO  >    IMAGE "SYS$SYSEXE:FOO.EXE"  >    QUALIFIER BAR > / > then adding it to the system DCLTABLES using:  > 3 > $ SET COMMAND/TABLE=SYS$LIBRARY:DCLTABLES.EXE -_$ 9 > /OUTPUT=SYS$COMMON:[SYSLIB]DCLTABLES.EXE NEWCOMMAND.CLD  >  > and  > - > $ INSTALL REPLACE SYS$LIBRARY:DCLTABLES.EXE  > E > So, I have several questions on this that I can't find answers for, 	 > namely:  > H > - all the examples I've seen of .cld files refer to "images", eg, .exe7 > files - will the CDU let me do this with a .com file?  > 9 > - if it will, I'm guessing my .cld file syntax will be:  >  > DEFINE VERB MYPROGRAM   >   IMAGE "@[DIR]MY_PROGRAM.COM" > G > as I don't have any qualifiers (just arguments).  I'm a bit unsure of I > the syntax, as it's a .com file, and the symbol defined earlier had the 0 > "@" in it - not sure if that's necessary here. > E > - if I do add this to the DCLTABLES, and some users have it already G > defined in their login scripts, will this conflict in any way?  If we F > have to change something in the future, which would take precedence? > G > - if for some reason I wanted to remove this entry from the DCLTABLES G > after adding it, how can I do that?  I've only found reference to SET E > COMMAND/DELETE, which appears to only operates on the local copy of   > DCLTABLES, not the global one. >  > Thanks for all suggestions.  >  > Glenn    Simpler just to add the line  ( > 'NEWPROGRAM :== "@[DIR]NEW_PROGRAM" '. to SYS$MANAGER:SYLOGIN.COM   Mike --  E --------------------------------------------------------------------- E Usual disclaimer: All opinions are mine alone, perhaps not even that. ? Mike Rechtman                            *rechtman@tzora.co.il* F Kibbutz Tzor'a.                          Voice (home): 972-2-9908337  B   "20% of a job takes 80% of the time, the rest takes another 80%"E ---------------------------------------------------------------------  -----BEGIN GEEK CODE BLOCK-----  Version: 3.1: GCM/CS d(-)pu s:+>:- a++ C++ U-- L-- W++ N++ K? w--- V+++$6 PS+ PE-- t 5? X- tv-- b+ DI+ D-- G e++ h--- r+++ y+++@ ------END GEEK CODE BLOCK------    ------------------------------  * Date: Mon, 9 Jan 2006 12:35:02 +0000 (UTC)< From: gartmann@nonsense.immunbio.mpg.de (Christoph Gartmann)8 Subject: Re: foreign commands, .com files, and dcltables) Message-ID: <dptl9m$bu0$1@news.BelWue.DE>   ] In article <1136702859.840797.313230@f14g2000cwb.googlegroups.com>, mabbuttg@yahoo.ca writes: , >I'm fairly new to OpenVMS, so bear with me: > E >A consultant implemented a process for us where several programs now B >call a .com file with arguments to complete some file operations. > H >To make this work, we had to add a foreign command symbol definition to5 >our users' login scripts, similar to 'NEWPROGRAM :==  >"@[DIR]NEW_PROGRAM" '.   K Usually there is a LOGIN.COM in each users home directory and a SYLOGIN.COM N for all the users on the whole system. In your case it would be best to modify just the latter.  G >- all the examples I've seen of .cld files refer to "images", eg, .exe 6 >files - will the CDU let me do this with a .com file?  0 No, as far as I'm aware of you'll need an image.  D >- if I do add this to the DCLTABLES, and some users have it alreadyF >defined in their login scripts, will this conflict in any way?  If weE >have to change something in the future, which would take precedence?   G Symbols defined in LOGIN.COM override the ones in the DCL-tables. So it > doesn't produce any error but you should know how things work.  F >- if for some reason I wanted to remove this entry from the DCLTABLESF >after adding it, how can I do that?  I've only found reference to SETD >COMMAND/DELETE, which appears to only operates on the local copy of >DCLTABLES, not the global one.   8 It will work with the system-wide command table as well.  M In General I consider it not a good idea to modify the system-wide DCL-tables E unless you know exactly what you are doing and why your are doing it. L It is a lot more easy to define a symbol in SYLOGIN.COM, it is more flexible and documents itself.    Regards,    Christoph Gartmann    --  E  Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452   ImmunbiologieI  Postfach 1169                 Internet: gartmann@immunbio dot mpg dot de   D-79011  Freiburg, Germany 9                http://www.immunbio.mpg.de/home/menue.html    ------------------------------  % Date: Mon, 09 Jan 2006 05:17:54 -0800 ( From: Jeff Cameron <roktsci@comcast.net>8 Subject: Re: foreign commands, .com files, and dcltables0 Message-ID: <BFE7A282.19FFA%roktsci@comcast.net>   On 1/7/06 10:47 PM, in articleJ 1136702859.840797.313230@f14g2000cwb.googlegroups.com, "mabbuttg@yahoo.ca" <mabbuttg@yahoo.ca> wrote:  - > I'm fairly new to OpenVMS, so bear with me:  > F > A consultant implemented a process for us where several programs nowC > call a .com file with arguments to complete some file operations.  > I > To make this work, we had to add a foreign command symbol definition to 6 > our users' login scripts, similar to 'NEWPROGRAM :== > "@[DIR]NEW_PROGRAM" '. > E > While we've gone through and added this symbol to what we're pretty C > sure are all the users who are affected by this, I was doing some H > research to see what could be done on a global level just to make sureD > (in our setup it doesn't matter if all users have the above symbolI > defined, as we have additional access restrictions on the programs that 4 > call it, and most users don't have prompt access). > I > What I've found is some info on modifying the global DCLTABLES, using a ' > .cld file, with the sample syntax of:  >  >  > DEFINE VERB FOO  >    IMAGE "SYS$SYSEXE:FOO.EXE"  >    QUALIFIER BAR > / > then adding it to the system DCLTABLES using:  > 3 > $ SET COMMAND/TABLE=SYS$LIBRARY:DCLTABLES.EXE -_$ 9 > /OUTPUT=SYS$COMMON:[SYSLIB]DCLTABLES.EXE NEWCOMMAND.CLD  >  > and  > - > $ INSTALL REPLACE SYS$LIBRARY:DCLTABLES.EXE  >  > E > So, I have several questions on this that I can't find answers for, 	 > namely:  > H > - all the examples I've seen of .cld files refer to "images", eg, .exe7 > files - will the CDU let me do this with a .com file?  > 9 > - if it will, I'm guessing my .cld file syntax will be:  >  > DEFINE VERB MYPROGRAM   >   IMAGE "@[DIR]MY_PROGRAM.COM" > G > as I don't have any qualifiers (just arguments).  I'm a bit unsure of I > the syntax, as it's a .com file, and the symbol defined earlier had the 0 > "@" in it - not sure if that's necessary here. >  > E > - if I do add this to the DCLTABLES, and some users have it already G > defined in their login scripts, will this conflict in any way?  If we F > have to change something in the future, which would take precedence? > G > - if for some reason I wanted to remove this entry from the DCLTABLES G > after adding it, how can I do that?  I've only found reference to SET E > COMMAND/DELETE, which appears to only operates on the local copy of   > DCLTABLES, not the global one. >  > Thanks for all suggestions.  >  > Glenn  > C As you have seen in other responses, the DCL tables and the command D definition utility does not allow you to define commands for commandH procedures. This opens a rather large security hole, but that aside, theK whole purpose of the DCL tables is to allow images to have a common command . parsing method with parameters and qualifiers.  K One other point; while you can change the actual system wide DCL tables, it L is typically not a good idea. This is because when you upgrade or patch VMS,H the kit my overwrite your changes. Instead (if you go the executable/CLDL route) that you include your SET COMMAND in the system wide login procedure.  L This is because each process that uses DCL copies the entire DCL tables intoH the context of the process when the process is created, so that when you@ execute the SET COMMAND it only changes the process' DCL tables.  F But if your whole purpose is to add a new global command for a commandG procedure, the best and most accepted method is to make a global symbol D definition in your system wide login procedure. One other point, forJ security reasons you would typically want to assign these file protections7 to any globally used command procedures to avoid hacks:   ; $SET SECURITY/PROT=(S:RE,O:RWED,G:RE,W:RE) <procedure-name>    Jeff   ------------------------------   Date: 8 Jan 2006 00:10:44 -0800 7 From: "greg.chabala@gmail.com" <greg.chabala@gmail.com> / Subject: Re: Free to a good home: OSF/1 manuals B Message-ID: <1136707843.962139.12110@g47g2000cwa.googlegroups.com>  . Ooooh, i want it! I'm living in La Grange, IL.   My email is bchabala2@yahoo.com    Let me know more details.    ------------------------------   Date: 9 Jan 2006 08:00:16 -0600 2 From: kaplow_r@encompasserve.org.mars (Bob Kaplow)/ Subject: Re: Free to a good home: OSF/1 manuals 3 Message-ID: <31q9NHknbKoG@eisner.encompasserve.org>   | In article <1136707843.962139.12110@g47g2000cwa.googlegroups.com>, "greg.chabala@gmail.com" <greg.chabala@gmail.com> writes:0 > Ooooh, i want it! I'm living in La Grange, IL.  K This discussion has been taken offline, but these manuals are spoken for. I J should have more stuff as I continue to clean up. More VMS than Tru64, butK who knows what's in those boxes that haven't been opened for 10-15 years...    --  O   Bob Kaplow   NAR # 18L   >>> To reply, there's no internet on Mars (yet)! <<< E Kaplow Klips & Baffle:	http://nira-rocketry.org/Document/MayJun00.pdf L     www.encompasserve.org/~kaplow_r/    www.nira-rocketry.org    www.nar.org  H Vulcans believe peace should not depend on force. -- Amanda, "Journey to Babel," stardate 3842.3    ------------------------------  * Date: Mon, 9 Jan 2006 17:26:14 +0000 (UTC). From: klewis@LUMINA.MITRE.ORG (Keith A. Lewis), Subject: Re: Java 1.4.2 installation problem. Message-ID: <dpu6bm$ein$1@newslocal.mitre.org>   "Stanley F. Quayle" <squayle@insight.rr.com> writes in article <43C188D1.5734.106AA896@localhost> dated Sun, 08 Jan 2006 21:49:05 -0500:F >I have just installed the latest Java from the HP web site (the 142p-+ >2 kit) and always get the following error:  >  >Case 1  >------ * >$ @SYS$COMMON:[SYSMGR]JAVA$142_SETUP FAST >$ java -version+ >Error: no `classic' JVM at `java$jvm_shr'.  >  >Case 2  >------ - >$ @SYS$COMMON:[SYSMGR]JAVA$142_SETUP CLASSIC  >$ java -version+ >Error: no `classic' JVM at `java$jvm_shr'.  >  >  >Any suggestions?   % Do you get the the following results?    $ dir java$jvm_shr  5 Directory SYS$COMMON:[JAVA$142.JRE.LIB.ALPHA.CLASSIC]    JAVA$JVM_SHR.EXE;1   Total of 1 file.  E If not, look for why not.  The setup script is supposed to point that & logical name to the correct directory.  0 --Keith Lewis              klewis {at} mitre.org> The above may not (yet) represent the opinions of my employer.   ------------------------------   Date: 9 Jan 2006 08:24:34 -0800 ; From: "vmsmangler@earthlink.net" <vmsmangler@earthlink.net>  Subject: Re: KVM Switches C Message-ID: <1136823874.919989.234000@g49g2000cwa.googlegroups.com>    Ken Fairfield wrote:! > vmsmangler@earthlink.net wrote:  > > William Webb wrote:  > > : > >>On 5 Jan 2006 13:26:22 -0800, vmsmangler@earthlink.net% > >><vmsmangler@earthlink.net> wrote:  > >> > >>>Galen wrote:  > >>> K > >>>>We have one kind of Belkin 2-port KVM switches here and another newer ( > >>>>switch from IOgear--model GCS612A. > >>>>K > >>>>With my particular LK461, both of them have the same minor problem if I > >>>>certain special keys are pressed repeatedly. The symptom is that it I > >>>>sort of freezes up until another special key is pressed. One of the J > >>>>keys with this problem happens to be PF3, which is a pain since it's. > >>>>used for searching in EDT-style editors. > >>>>K > >>>>I don't have another LK461 immediately available to determine whether K > >>>>the problem is in  the keyboard itself, or in the switches. I'll have B > >>>>to borrow one from another system when nobody's looking. :-) > >>> K > >>>Thanks for the information. I checked iogear's website and also called  > >>>their technical people.K > >>>They claim to only support the 102 key keyboard so I am surprised that  > >>>is supports a true DEC L > >>>keyboard, e.g. LK461. From what I have determined from all responses is > >>>that PC's keyboardsH > >>>are scanset 2 while DEC keyboards are scanset 3. My 2 DPW 433/500au > >>>workstations all haveJ > >>>keyboard detection problems during power-up testing by SRM. They also > >>>have the error that% > >>>scan code 0xfa is not supported.  > >>> 	 > >>>Bill  > >>>   > >>>Retired VMS Systems Manager > >>>  > >>>  > >>I > >>I seem to recall that there is much anecdocal data on which KVMs work A > >>with LK-style keyboards in the comp.os.vms historical record.  > >>
 > >>WWWebb > >>--G > >>NOTE: This email address is only used for noncommerical VMS-related  > >>correspondence. G > >>All unsolicited commercial email will be deemed to be a request for < > >>services pursuant to the terms and conditions located at' > >>http://bellsouthpwp.net/w/e/webbww/  > >  > > A > > I have checked the historical records for KVM switches, LK46x I > > keyboards, etc. Wow! Lots of information and somewhat conflicting but  > > very usefull.  > > K > > I am curious if Ken Fairfield ever got the Belkin switch to work with a : > > true Alpha running VMS and using a LK46 type keyboard. >  > Well, yes, with conditions...  > B > First as a reminder, I use an LK462 (black version of an LK461),@ > an older-DEC-style Logitech 3-button mouse, and a "modern" LCDD > Samsung display, between a DPW600au and an IBM T41 laptop-via-port; > replicator (to get to the external keyboard/mouse/video).  > > > I had some initial problems with the DPW600au when hooked up> > through the Belkin.  I finally determined that the mouse and> > keyboard connections were "flaky"; the plugs from the Belkin: > cables didn't seat firmly in the receptacles on the DPW.< > I solved that by trimming about 1mm of the insulation from: > the ground/barrel of the mouse and keyboard plugs.  That > solved the problem.  > ? > I also seem to get what would appear to be discharge flickers = > on the monitor after the DPW has been on a while.  It has a B > "standard" Elsa Gloria video card, so you'd think there wouldn'tA > be any connector compatibility issues.  I have it on my list to = > go back and see if I can do something similar to what I did D > with the keyboard & mouse connections to correct it, but it hasn't+ > been enough of a bother to deal with yet.  > @ > I have absolutely no problem using the laptop.  I also have noA > problem on the DPW console, after power-up or shutdown (someone 0 > had mentioned a problem along those lines...). > ? > With both the laptop and the DPW, ALL the LK keys function as B > expected/desired (with either Exceed or KEA and suitable mapping > tables on the laptop). > F > Here's the big "your mileage may vary": I rarely have both computers@ > on at the same time, so I rarely exercise the actual switchingF > between computer function (Alt-F19, or Shift-F19, on an LK).  I haveE > done test and verified it works.  I have verified that, contrary to C > the documentation, the switch WON'T go into continuous scan mode, A > something John Malmberg mentioned.  [I have managed to get into D > situations where the active computer (the one displaying) wouldn'tD > respond to the keyboard, but I think I was specifically banging onC > the F19 thingy in a way that confused the switch, or confused the C > computer...]  But I believe for routine use, the switch functions  > quite adequately.  >  >  > Some commentary: > --------------- B > I greatly prefer this "simple" Belkin over the Rose Mini-Vista IG > previously used.  The difference is that the Belkin, when it switches I > (every time it switches), resends the keyboard initialization sequence. H > That means (a) there is a slight delay before you can type anything toC > the just-switched-to computer, but (b) the keyboard is put in the G > correct scan mode for that computer.  In addition, the Belkin appears D > to just pass through any and all key sequences, not just a (semi-) > limited set it knows about.  > F > The Rose, on the other hand, insists on (a) maintaining the keyboardD > in scan mode 2, and (b) translating key sequences to the scan modeG > the computer wants, scan mode 3 for the DPW.  But for the DPW, that's E > a Catch-22 since at power-up, the console doesn't send the keyboard F > reset sequence (that VMS eventually does), so the Rose tries to talkF > to it in scan mode 2 instead.  Ugh.  You can program the port on theC > Rose to talk scan mode 3 (and there's a different special setting G > required for the mouse, IIRC), but when the Rose loses power, because H > both computers are off, it loses its programming!  The work-around wasE > to put a Rose-supplied external power supply between the switch and G > one of the two input keyboard lines.  Finally, there were a number of I > LK key sequences that the Rose simply didn't know about and so wouldn't H > do the translation on.  For the all the time I used the Rose, I neededD > TWO KEYBOARDS: one connected through the KVM that I'd use with the@ > laptop, and the second directly plugged into the DPW.  Sigh... > G > I'm told by Rose that some of their more expensive models (but older) F > work correctly with LK keyboards in scan mode 3.  I'm also told thatH > some Raritan models, possibly with additional required "options", alsoF > work with LK keyboards and scan mode 3.  In both cases, the completeF > solution is MUCH MORE EXPENSIVE and not necessarily more functional. > I > In summary, the "cheap" commodity KVMs, while not officially supporting E > an LK keyboard, seem to work pretty well, at least the Belkin does. G > A bonus with the particular Belkin John recommended, the F1DL102P, is H > that it really saves real estate by having the two computer cable sets. > built-in/preconnected internally to the box. >  >      -Ken  > --8 > I don't speak for Intel, Intel doesn't speak for me... >  > Ken Fairfield # > D1C Automation VMS System Support $ > who:   kenneth dot h dot fairfield > where: intel dot com  @ Thank you Ken for your great reply. You have done very extensiveE testing with the Belkin switch which confirms my thoughts about which D switch to purchase. Based upon your results I have decided to obtainC the Belkin F1DJ102P. This switch does not support audio but it does D have a button to switch systems. Hot Key switching is also supported0 but I really like the idea of a separate button.  C I will post the results of my testing this switch as soon as I can.   0 Thanks to all who have responded to this thread.  
 Bill LaCounte  Retired VMS Systems Manager    ------------------------------   Date: 8 Jan 2006 19:32:41 -0800 3 From: "Joshua Lehrer" <usenet_vms@lehrerfamily.com> $ Subject: Re: more lbr$flush problemsC Message-ID: <1136666672.585128.119690@g43g2000cwa.googlegroups.com>    JF Mezei wrote:  > Joshua Lehrer wrote:F > > However, I have discovered that calling lbr$flush between calls to: > > lbr$put_record causes the library to become corrupted. > H > Also, you can't call lbr$close immediatly after lbr$flush. You need toO > do something to the library in between the two, otherwise you get corruption.   F I know.  I reported this to HP a while back, and they sent me a patch. Further, it is fixed under 8.x.    > G > The easiest solution would be to increase the virtual memory quota of  > your process (pgfilquota).  D Increasing the virtual memory quota does nothing to make more 32-bitB virtual address space available.  Your process gets only 2-gigs ofF non-shared 32-bit virtual address space, P0 and P1.  The lbr$ routinesE can only use P0 memory, and thus are limited to 1gig.  This is shared A with the rest of my program's use of malloc/new, vmzones, images,  global data, etc...   D Again, I see no way to create a module with more than 1 gig of data.E Further, if you have only 200 megs of 32-bit P0 virtual address space G available (due to the other 800+ megs already being used), then you can 2 not make a module with more than 200 megs of data.  G I see two fixes that VMS could employ.  First, allow calls to lbr$flush F between calls to lbr$put_record.  Or, second, update lbr$put_record toA copy the line into 64-bit address space instead of 32-bit address F space.  Then, your only limitation would be your memory quotas and not5 the amount of 32-bit virtual address space remaining.    -josh    ------------------------------  % Date: Mon, 09 Jan 2006 04:09:17 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> $ Subject: Re: more lbr$flush problems, Message-ID: <43C22801.D5551614@teksavvy.com>   Joshua Lehrer wrote:F > Again, I see no way to create a module with more than 1 gig of data.  G Ok, I truly had not considered the creation of such large modules. That E is a hell of a lot of data to put in a library especially if you have  multiple modules that size.   D I get the impression that the whole library subsystem was abandonnedE some time ago and they just keep it running. They never completed the H documentation for it, as an example (for instance documenting the formatA of history records), or documenting how to get the same data that % LIB/DIR/FULL gives you in the header.    ------------------------------   Date: 9 Jan 2006 05:15:01 -0800 3 From: "Joshua Lehrer" <usenet_vms@lehrerfamily.com> $ Subject: Re: more lbr$flush problemsC Message-ID: <1136812501.248358.162070@g49g2000cwa.googlegroups.com>    JF Mezei wrote:  > Joshua Lehrer wrote:H > > Again, I see no way to create a module with more than 1 gig of data. > I > Ok, I truly had not considered the creation of such large modules. That G > is a hell of a lot of data to put in a library especially if you have  > multiple modules that size.  >   G My problem is that the modules may not even be that large.  If you have D little 32-bit virtual memory remaining, you can create only a little@ module.  The size of a module is limited by the amount of 32-bitG virtual address space remaining and, while running some of my programs, * 32-bit address space is a scarce resource.  C It would be nice if someone could fix the known lbr$flush bugs, add F 64-bit support to the lbr$ routines, and improve the internal cacheing to use P2 instead of P0.  
 joshua lehrer    ------------------------------  # Date: Mon, 09 Jan 2006 17:37:12 GMT & From: John Reagan <john.reagan@hp.com>$ Subject: Re: more lbr$flush problems0 Message-ID: <cbxwf.1384$%O.420@news.cpqcorp.net>   Joshua Lehrer wrote:   > E > It would be nice if someone could fix the known lbr$flush bugs, add H > 64-bit support to the lbr$ routines, and improve the internal cacheing > to use P2 instead of P0. >  > joshua lehrer  >   - Can you submit a real problem report on that?   G The person who supports the LBR$ routines doesn't read this newsgroup.  C Plus the official report lets the management know that we're doing  H support issues.  (If you don't have a support contract, let me know and & I'll forward the whole thing along...)   --   John Reagan / HP Pascal/{A|I}MACRO for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------   Date: 9 Jan 2006 09:15:27 -0800 < From: "Hein RMS van den Heuvel" <heinvandenheuvel@gmail.com>$ Subject: Re: more lbr$flush problemsC Message-ID: <1136826927.709727.176180@z14g2000cwz.googlegroups.com>    Joshua Lehrer wrote: > F > Again, I see no way to create a module with more than 1 gig of data.    $ And IMHO that maybe is just as well.2 Something has to stop you and make you reconsider.E LBR is highly appropriate for lots of little files, much more so than F sticking them into little files wach with their file header, directory0 entry and cluster size space allocation roundup.B But there comes a point where you probably just store you files as file. G My recommendation for that point is 10 MB, but before reading this note ! it might have been 100 KB or 1MB. F If you still need the LBR structure to make teh module work similar toD others, then i woudl recommend an indirection level where the module  points to the real file somehow.  G Please help us understand why a 100+ MB module in LBR file is the right % approach to solve a business problem.    Regards, Hein.    ------------------------------   Date: 9 Jan 2006 10:44:02 -0800 3 From: "Joshua Lehrer" <usenet_vms@lehrerfamily.com> $ Subject: Re: more lbr$flush problemsC Message-ID: <1136832242.810117.141390@g47g2000cwa.googlegroups.com>   / That's not really my issue.  I have two issues:   A 1] Any module I create is limited in size by the amount of 32-bit G virtual address space available.  An easy fix for this would be to have @ the lbr$put_record routine cache the lines in P2 rather than P0.  C 2] lbr$flush does not work as advertised, the documentation implies D that lbr$flush can be called between calls to lbr$put_record, but it2 can not.  It would be nice if this worked as well.   joshua   ------------------------------   Date: 9 Jan 2006 07:35:32 -0600 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) 2 Subject: Re: Need model number for Infoserver-10003 Message-ID: <6bYEuLd+QeMn@eisner.encompasserve.org>   k In article <2qSdnUKIyOzh_lzenZ2dnUVZ_tWdnZ2d@comcast.com>, Bob Blunt <RobertDOTblunt@digitalDOTcom> writes:  > J > Look at the bottom of your Infoserver 1000.  Hopefully it should have a K > sticker with TWO entries for the model number and the 70-30343-03 A02 is  H > the bottom one and in parenthesis.  As others have stated, the number J > you need will be of the format SEADx-xx.  The 70-30343-03 is really the K > "option" part number of the daughter card that plugs onto the Infoserver  I > motherboard that provides your console, ethernet and power connections  H > (in the case of the 70-30343-03, this is the AUI variant).  There are J > also variants of those daughter cards options that are Thinwire and one   > that plugs into the Infotower. > ) > 70-30343-01  SEADB, standalone thinwire $ > 70-30343-03  SEADC, standalone AUI > 70-30343-04  Infotower  F    Thanks.  The sticker must have been misprinted, there is no SEAD...?    on it.  I've determined the model number from earlier posts.    ------------------------------  % Date: Mon, 09 Jan 2006 07:53:52 +0100 3 From: Michael Unger <spam.to.unger@spamgourmet.com> * Subject: Re: old EZ disks with latest VMS?, Message-ID: <42eh62F1ilokvU1@individual.net>  ' On 2006-01-08 22:49, "Bob Blunt" wrote:    > [...]  > H > [...]  In a nutshell, they have a built-in battery and RZ25 long-term J > retention backup but the batteries have died and expired.  All sorts of A > weird things happen and the replacement battery packs are VERY   > expensive.  [...]    ^^^^^^^^^   E What type of "battery" is used in these SSDs? NiCd, NiMH, LiIon, ...?    Michael    --  ; Real names enhance the probability of getting real answers. 5 My e-mail account at DECUS Munich is no longer valid.    ------------------------------   Date: 9 Jan 2006 07:58:06 -0600 2 From: kaplow_r@encompasserve.org.mars (Bob Kaplow)* Subject: Re: old EZ disks with latest VMS?3 Message-ID: <rLlX2Ahrbgjm@eisner.encompasserve.org>   b In article <42eh62F1ilokvU1@individual.net>, Michael Unger <spam.to.unger@spamgourmet.com> writes:G > What type of "battery" is used in these SSDs? NiCd, NiMH, LiIon, ...?   J This is OLD technology, so it's unlikely to be either NiMH or LoIon. Maybe$ nicad, more likely sealed lead acid.  I For modern environments, it would make more sense to forget about SSD and G just increase main memory and cache the data internally. Or use storage L controllers that have big internal cache. Either way, current usage patternsG define what is in memory and what is not, rather than predefined rules.    --  O   Bob Kaplow   NAR # 18L   >>> To reply, there's no internet on Mars (yet)! <<< E Kaplow Klips & Baffle:	http://nira-rocketry.org/Document/MayJun00.pdf L     www.encompasserve.org/~kaplow_r/    www.nira-rocketry.org    www.nar.org  H Vulcans believe peace should not depend on force. -- Amanda, "Journey to Babel," stardate 3842.3    ------------------------------   Date: 7 Jan 2006 23:25:43 -0800 , From: "rcyoung" <rcyoung@aliconsultants.com>6 Subject: Recommendations for 4mm tape on 4000/600 Vax?C Message-ID: <1136668153.002383.273540@z14g2000cwz.googlegroups.com>   ? Looking for recommendations for a 4mm tape (6-8 tape autoloader F preferred) unit that  will work with a 4000/600 running OpenVMS 7.1. ID have looked through the "old" SOC catalogs, and the TLZ7L/TLZ9L lookF most promising, but I ned to verify they will work on the Vax 4000/600D w/ 7.1 or higher. Most of the SOCs mentioning these drives "predate"G the VMS 7 .  A couple mention Vax VMS 6.1 would need an enhancement kit 1 added, but say nothing about higher VMS versions.    Suggestions?   ------------------------------  $ Date: Mon, 9 Jan 2006 11:49:47 -0500C From: "David Turner, Island Computers US Corp" <dbturner@icusc.com> : Subject: Re: Recommendations for 4mm tape on 4000/600 Vax?8 Message-ID: <Jmwwf.15235$__4.667@bignews1.bellsouth.net>    I do have some of the following:  C Compaq DDS2 4/8GB Autoloader (4 cassette with a 12 cassette option)    Table top $500   David    --     David B Turner Island Computers US Corp 2700 Gregory St, Suite 180 Savannah GA 31404  Tel: 912 447 6622 X201 Cell: 912 447 6622 X252  Fax: 912 201 0402  Email: dbturner@icusc.com  Web: http://www.islandco.com% ===================================== < All orders are subject to the following terms and conditions. of sale. These should be read before ordering.% http://www.islandco.com/warranty.html   7 "rcyoung" <rcyoung@aliconsultants.com> wrote in message = news:1136668153.002383.273540@z14g2000cwz.googlegroups.com... A > Looking for recommendations for a 4mm tape (6-8 tape autoloader H > preferred) unit that  will work with a 4000/600 running OpenVMS 7.1. IF > have looked through the "old" SOC catalogs, and the TLZ7L/TLZ9L lookH > most promising, but I ned to verify they will work on the Vax 4000/600F > w/ 7.1 or higher. Most of the SOCs mentioning these drives "predate"I > the VMS 7 .  A couple mention Vax VMS 6.1 would need an enhancement kit 3 > added, but say nothing about higher VMS versions.  >  > Suggestions? >    ------------------------------   Date: 9 Jan 2006 08:01:53 -0800 ( From: "Rich Jordan" <jordan@ccs4vms.com>* Subject: Re: Round internal SCSI cables ??C Message-ID: <1136822513.017147.217950@z14g2000cwz.googlegroups.com>    Jilly wrote:F > Anyone ever seen round internal SCSI cables al what you can find forN > internal IDE cables?  I need to replace the 17-04474-01 & 17-04009-01 cablesG > on my PWS 600au so that I can fit a few more options inside the case.  >  > Jilly   C I have no experience with this vendor or products, but they seem to  have a good selection...  ' http://www.xpcgear.com/scsirouncab.html    Ditto with this one:  9 http://www.cablemakers.com/round_internal_scsi_cables.htm    Rich CCS    ------------------------------  $ Date: Mon, 9 Jan 2006 21:02:18 +08003 From: "Richard Maher" <maher_rj@hotspamnotmail.com> ) Subject: Re: UNIX shm* functions for VMS? 0 Message-ID: <dptmtl$uk$1@news-02.connect.com.au>  	 Hi Keith,   E Where've you been? I (and others in venues such as de.comp.os.vms who K thought you were Mark Twaining it) have presumed that you were convalescing 7 in some sort of Sharon-esque Adelstoff induced coma :-)   K Please reply if you still have at least a passing interest in VMS<->W2K TIP  2PC functionality.  " Good to have you back, regardless.   Cheers Richard Maher  = "Keith Cayemberg" <keith.cayemberg@arcor.de> wrote in message < news:43c233da$0$20772$9b4e6d93@newsread4.arcor-online.net... > Steven M. Schweda wrote:L > >    I seek VMS-compatible code (ideally, already written and working) forK > > the UNIX shared memory functions, like shmat(), shmctl(), shmget(), and  > > friends. > >  > >    I found:  > > H > >       http://vmsone.com/~decuslib/vmssig/vmslt02a/vu/gnv-crtlsup.htm > > 5 > > which looked good but offered only a broken link:  > > L > >       http://cvs.sourceforge.net/cgi-bin/cvsweb.cgi/crtlsup/?cvsroot=gnv > > : > >    Anyone with more than that is welcome to inform me. > > L > > ------------------------------------------------------------------------ > > 8 > >    Steven M. Schweda               (+1) 651-699-98187 > >    382 South Warwick Street        sms@antinode-org ! > >    Saint Paul  MN  55105-2547  >  > 
 > Try this...  > , > http://cvs.sourceforge.net/viewcvs.py/gnv/ > J > your broken link still exists in the Internet Archive (Wayback Machine). >  > _ http://web.archive.org/web/*/http://cvs.sourceforge.net/cgi-bin/cvsweb.cgi/crtlsup/?cvsroot=gnv  > B > It is however just a further link to the Sourceforge site above. > 	 > Cheers!  >  > Keith    ------------------------------  % Date: Mon, 09 Jan 2006 14:49:22 +0100 ( From: JOUKJ <joukj@hrem.nano.tudelft.nl>) Subject: Re: UNIX shm* functions for VMS? 3 Message-ID: <43C269E2.2030609@hrem.nano.tudelft.nl>    Steven M. Schweda wrote:J >    I seek VMS-compatible code (ideally, already written and working) forI > the UNIX shared memory functions, like shmat(), shmctl(), shmget(), and 
 > friends. > ? On my system (OpenVMS8.2 for alpha with DecWindows 1.5) I find  & definitions of shmat() and shmctl() in,    SYS$COMMON:[DECW$INCLUDE.EXTENSIONS]SHM.HG the definition of shmget is obviously missing. But programs containing  D it could be linked against the DECW-libraries. The single program I 8 tested (the fzort demo of xlockmore) worked in this way.                         Jouk   ------------------------------   Date: 8 Jan 2006 17:36:23 -0800 . From: "Schnootling" <chuckmoore55@hotmail.com>0 Subject: Re: Using VT420 as Alpha System ConsoleC Message-ID: <1136770583.694076.236640@g43g2000cwa.googlegroups.com>    Hello Steven and Everybody,   8 I've done some investigation and here's what I've found:B 1) I originally had the connection built to connect a VT420 to the serial port on a PC., 2) The connection is constructed as follows:8 VT420<->MMJ -to- RJ12 male<->RJ12 plug -to- Sub-D 25-pinC adapter<->25-pin gender changer<-> Sub-D 25-pin (fully-pinned) male  -to- 9-pin female.  G I'm pretty sure the RJ12 is a RJ12 and not a RJ11 because the RJ12 plug E has 6 wires in it. I took the Sub-D 25-pin adapter apart but couldn't 4 see very well (inside) how everything was connected.  9 The used pins on the Sub-D 25-pin adapter are 2-8 and 20.   6 I haven't had a chance to do the loopback thingee yet.  B Maybe the shortest-path to getting the VT420 useable is to order a9 DEC-whatever that'll take RJ12 straight into the Alpha ??    Chuck   & P.S. Got my connector information from+ http://www.connectworld.net/connectors.html    ------------------------------  % Date: Mon, 09 Jan 2006 11:49:28 +0200 4 From: Mike Rechtman <michael.rechtman.nospam@hp.com>0 Subject: Re: Using VT420 as Alpha System Console% Message-ID: <43C24DC8.BF6734E@hp.com>    Schnootling wrote: >  > Hello Steven and Everybody,  > : > I've done some investigation and here's what I've found:D > 1) I originally had the connection built to connect a VT420 to the > serial port on a PC.. > 2) The connection is constructed as follows:: > VT420<->MMJ -to- RJ12 male<->RJ12 plug -to- Sub-D 25-pinE > adapter<->25-pin gender changer<-> Sub-D 25-pin (fully-pinned) male  > -to- 9-pin female. > I > I'm pretty sure the RJ12 is a RJ12 and not a RJ11 because the RJ12 plug G > has 6 wires in it. I took the Sub-D 25-pin adapter apart but couldn't 6 > see very well (inside) how everything was connected. > ; > The used pins on the Sub-D 25-pin adapter are 2-8 and 20.  > 8 > I haven't had a chance to do the loopback thingee yet. > D > Maybe the shortest-path to getting the VT420 useable is to order a; > DEC-whatever that'll take RJ12 straight into the Alpha ??  >  > Chuck  >   ( > P.S. Got my connector information from- > http://www.connectworld.net/connectors.html   E 1. Try to get hold of (or construct) an MMJ loopback plug. This is an E MMJ plug with pairs of contacts shorted: 1<->6  2<->5  3<->4. Plug it E into the MMJ socket you are using on the VT420, and check whether the D screen "echoes" whatever you type. If it does not - you have a VT4201 problem; Solve it before anything else will work. C 2. The simplest way of connecting an MMJ socket to a 9-pin DB9 plug G would be an MMJ <-> MMJ cable BC16E-??, and MMJ to DB9 adapter (H8571-J  ??) Try to keep it simple!  = See the OpenVMS FAQ, section 14.27 for details, pinouts, etc.      Mike --  E --------------------------------------------------------------------- E Usual disclaimer: All opinions are mine alone, perhaps not even that. ? Mike Rechtman                            *rechtman@tzora.co.il* F Kibbutz Tzor'a.                          Voice (home): 972-2-9908337  B   "20% of a job takes 80% of the time, the rest takes another 80%"E ---------------------------------------------------------------------  -----BEGIN GEEK CODE BLOCK-----  Version: 3.1: GCM/CS d(-)pu s:+>:- a++ C++ U-- L-- W++ N++ K? w--- V+++$6 PS+ PE-- t 5? X- tv-- b+ DI+ D-- G e++ h--- r+++ y+++@ ------END GEEK CODE BLOCK------    ------------------------------  $ Date: Mon, 9 Jan 2006 10:49:02 -0500- From: William Webb <william.w.webb@gmail.com> 0 Subject: Re: Using VT420 as Alpha System ConsoleI Message-ID: <8660a3a10601090749l1ce7c756s78bd2bfef41d0ea5@mail.gmail.com>   K On 8 Jan 2006 17:36:23 -0800, Schnootling <chuckmoore55@hotmail.com> wrote:  > Hello Steven and Everybody,  > : > I've done some investigation and here's what I've found:D > 1) I originally had the connection built to connect a VT420 to the > serial port on a PC.. > 2) The connection is constructed as follows:: > VT420<->MMJ -to- RJ12 male<->RJ12 plug -to- Sub-D 25-pinE > adapter<->25-pin gender changer<-> Sub-D 25-pin (fully-pinned) male  > -to- 9-pin female. > I > I'm pretty sure the RJ12 is a RJ12 and not a RJ11 because the RJ12 plug G > has 6 wires in it. I took the Sub-D 25-pin adapter apart but couldn't 6 > see very well (inside) how everything was connected. > ; > The used pins on the Sub-D 25-pin adapter are 2-8 and 20.  > 8 > I haven't had a chance to do the loopback thingee yet. > D > Maybe the shortest-path to getting the VT420 useable is to order a; > DEC-whatever that'll take RJ12 straight into the Alpha ??  >  > Chuck  > ( > P.S. Got my connector information from- > http://www.connectworld.net/connectors.html  >  >   L 420's got MMJ serial ports, right?  Alpha's got DB9 Male serial ports, righ= t?  D The shortest (or should I say least complicated) path to you want isF to use a BC16E cable (MMJ on both ends) and an H8571-J (MMJ =3D> DB9F)	 connector    HTH.   WWWebb --C NOTE: This email address is only used for noncommerical VMS-related  correspondence. C All unsolicited commercial email will be deemed to be a request for 8 services pursuant to the terms and conditions located at# http://bellsouthpwp.net/w/e/webbww/    ------------------------------  % Date: Mon, 09 Jan 2006 03:10:47 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 4 Subject: Re: WVNETcluster uptime reaches 10 years...+ Message-ID: <43C21A4F.841725D@teksavvy.com>    Bill Todd wrote:J > Most folks would be wrong, then:  a central lock manager can handle thatF > just fine (and does in quite a few commercially-successful systems -/ > SANergy having been one of the earlier ones).   D In the case of a non VMS cluster, considering how much Unix tends toD cache file data, do these 3rd party clustering systems still provideH cache coherence on node1 when node2 has just written the blocks to disk,G  so that node1's app reads the updated blocks from disk insteads of its  cache ?   F A few years ago, I had been told that the Solaris version of PMDF usedE files as locks so that the SMTP server would know when not to deposit F new messages because the POP or IMAP applications on another node wereG accessing that account at the moment. (with a deamon that would cleanup G stray lock files that were left when a process ended abnormally without & cleaning up (deleting) its lock files.  ; What VMS does is make thing far more elegant and efficient.     H But VMS still lacks the ability to notify an application when a file has# just been deposited in a directory.    ------------------------------  % Date: Mon, 09 Jan 2006 13:28:06 -0500 ( From: Bill Todd <billtodd@metrocast.net>4 Subject: Re: WVNETcluster uptime reaches 10 years...G Message-ID: <pJWdnZSWF7ClNl_enZ2dnUVZ_s6dnZ2d@metrocastcablevision.com>    JF Mezei wrote:  > Bill Todd wrote: > J >>Most folks would be wrong, then:  a central lock manager can handle thatF >>just fine (and does in quite a few commercially-successful systems -/ >>SANergy having been one of the earlier ones).  >  > F > In the case of a non VMS cluster, considering how much Unix tends toF > cache file data, do these 3rd party clustering systems still provideJ > cache coherence on node1 when node2 has just written the blocks to disk,I >  so that node1's app reads the updated blocks from disk insteads of its 	 > cache ?   I Of course they do:  that's a large part of what lock management enforces  B (and it enforces it whether or not the data has yet actually been I written to disk, though sometimes by suppressing distributed caching via  I specific or general invalidation - as, of course, VMS also does - rather  G than by implementing a full-fledged distributed cache as, e.g., Oracle   does).   - bill   ------------------------------   End of INFO-VAX 2006.018 ************************