1 INFO-VAX	Wed, 16 Oct 2002	Volume 2002 : Issue 571       Contents:2 Re: $ RECALL/SEARCH and editing long command lines2 Re: $ RECALL/SEARCH and editing long command lines2 Re: $ RECALL/SEARCH and editing long command lines2 Re: $ RECALL/SEARCH and editing long command lines" 7.2-1 and 7.3-1 NFS and block size& Re: 7.2-1 and 7.3-1 NFS and block size% Re: Adaptec 39160 (KZPEA) & OpenVMS ? C Re: Anyone else having problems with Carepaqs and hardware support? + Re: Are there TK50 images for VMS avalable? * Re: Booting a VMS cluster node using BOOTP Re: CTRL-Y and VMS BACKUP  Re: CTRL-Y and VMS BACKUP  Re: CTRL-Y and VMS BACKUP  Re: DLM questions  Re: Hobbyist kit Contents $ Re: HoPping along with two left feet/ Re: How hard is this for HP to do?  Clustering? 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 RE: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 RE: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 RE: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"!   Re: Java performance on open VMS  Re: Java performance on open VMS Re: Life after VMS?  Re: Life after VMS?  Re: Life after VMS?  Re: Life after VMS?  Re: Life after VMS?  Re: Life after VMS?  Re: Life after VMS?  Re: Life after VMS?  Re: Life after VMS?  Re: Life after VMS?  Re: Life after VMS?  RE: Life after VMS?  Re: Life after VMS?  Re: Life after VMS?  Re: Life after VMS?  Re: Life after VMS?  Re: Life after VMS? ' Re: Ncurses porting and cbreak(ed) read  Re: New to VMS RE: New to VMS Re: Old DECC for VAX Re: Old DECC for VAX Re: OpenVMS is strategic Re: OpenVMS is strategic+ Re: Originator of detached process from DCL * Re: OT Humour - If Microsoft made toasters* RE: OT Humour - If Microsoft made toasters* Re: OT Humour - If Microsoft made toasters* Re: OT Humour - If Microsoft made toasters) Re: OVMS Concurrent user (update license)  Re: PeopleSoft on VMS  Re: PeopleSoft on VMS # Question about f$getsyi and Itanium ' Re: Question about f$getsyi and Itanium ' Re: Question about f$getsyi and Itanium  Re: singular and plural of VAX" Re: store .MOV file on a VMS disk?" stuck process after v7.3-1 upgrade
 Tape Copy? Re: Tape Copy?0 Re: TCPIP, sharing sockets via $QIO (fixed yet?)) Re: TCPIP,VMS 7.2.1 Alpha, license issue? ) Re: TCPIP,VMS 7.2.1 Alpha, license issue? ' Re: Telnet problem after 7.3-1 Upgrade. . Re: Unix file descriptors and VMS I/O channels. Re: Unix file descriptors and VMS I/O channels. Re: Unix file descriptors and VMS I/O channels. Re: Unix file descriptors and VMS I/O channels Re: VMS 5.3 disk installation 0 VMS 7.3-1 breaks clustering using memory channel RE: VMS 7.3-1 File Corruption 7 Re: VMS MAIL: COMPRESS _increases_ the size of the file ! Re: Why is LTA0: offline mounted? ! Re: Why is LTA0: offline mounted? ! Re: Why is LTA0: offline mounted? ! Re: Why is LTA0: offline mounted? 3 Re: Why slow system when using logical search list? # Re: ZDNet says you're dead by 2004! # Re: ZDNet says you're dead by 2004!   F ----------------------------------------------------------------------  # Date: Tue, 15 Oct 2002 20:14:26 GMT " From: Guy Peleg <guy.peleg@hp.com>; Subject: Re: $ RECALL/SEARCH and editing long command lines & Message-ID: <3DAC7547.C96C8C63@hp.com>  G The V7.3-1 DCL kit also contains new item codes for F$GETSYI to provide  information about memory usage' like TOTAL_MEMORY,USED_MEMORY and more.   N I do not plan to backport RECALL/SEARCH but if there is high enough demand you never know.....   M The biggest thing we (DCL engineering team) are currently working on is EDCL.  EDCL will increaseJ the command line limit to 4KB and to 8KB using the hyphen sign. This means bigger symbols bigger commandsN better ODS-5 support and bigger and better everything. EDCL will ship with the next OpenVMS version.  A Editing longer lines will be investigated once EDCL work is over.    As usual feedback is welcome.    Regards,  	 Guy Peleg  OpenVMS Engineering    Simon Clubley wrote:  L > The DCL ECO just released for V7.3-1 includes the command $ recall/search. > B > Are there any plans to extend this to previous versions of VMS ? > I > On a related note, are there any plans yet to extend DCL to add support 4 > for editing lines longer than the terminal width ? >  > Simon. >  > --= > Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP - > Microsoft: The Lada of the computing world.    ------------------------------  # Date: Tue, 15 Oct 2002 20:14:58 GMT " From: Guy Peleg <guy.peleg@hp.com>; Subject: Re: $ RECALL/SEARCH and editing long command lines & Message-ID: <3DAC756A.2C9B0A71@hp.com>  G The V7.3-1 DCL kit also contains new item codes for F$GETSYI to provide  information about memory usage' like TOTAL_MEMORY,USED_MEMORY and more.   N I do not plan to backport RECALL/SEARCH but if there is high enough demand you never know.....   M The biggest thing we (DCL engineering team) are currently working on is EDCL.  EDCL will increaseJ the command line limit to 4KB and to 8KB using the hyphen sign. This means bigger symbols bigger commandsN better ODS-5 support and bigger and better everything. EDCL will ship with the next OpenVMS version.  A Editing longer lines will be investigated once EDCL work is over.    As usual feedback is welcome.    Regards,  	 Guy Peleg  OpenVMS Engineering      >    Simon Clubley wrote:  L > The DCL ECO just released for V7.3-1 includes the command $ recall/search. > B > Are there any plans to extend this to previous versions of VMS ? > I > On a related note, are there any plans yet to extend DCL to add support 4 > for editing lines longer than the terminal width ? >  > Simon. >  > --= > Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP - > Microsoft: The Lada of the computing world.    ------------------------------    Date: 15 Oct 2002 15:41:06 -0600- From: koehler@encompasserve.org (Bob Koehler) ; Subject: Re: $ RECALL/SEARCH and editing long command lines 3 Message-ID: <j$2+hBxl9G09@eisner.encompasserve.org>   x In article <E$agPp00Nqxv@eisner.encompasserve.org>, clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) writes:L > The DCL ECO just released for V7.3-1 includes the command $ recall/search. > B > Are there any plans to extend this to previous versions of VMS ?  A    You want the new feature, that's what the new software is.  So A    DEC/Compaq/HP numbers the software release?  Either you change     the software or you don't.    ------------------------------  % Date: Wed, 16 Oct 2002 00:40:27 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> ; Subject: Re: $ RECALL/SEARCH and editing long command lines , Message-ID: <3DACEDAE.6DA0BE95@videotron.ca>   Guy Peleg wrote:P > I do not plan to backport RECALL/SEARCH but if there is high enough demand you > never know.....   4 I *demand* RECALL/SEARCH be available on 7.2 on VAX.   no, sorry...  6 I **DEMAND** RECALL/SEARCH be available on 7.2 on VAX.   ----------  F Is that enough demand ? or need I repeat the line 5000 times ? :-) ;-)  M Seriously though, out of curiosity, the RECALL/SEARCH was suggested here some I time ago. Is it possible that the implementation is the result of the all J mighty VMS engineers having seen the suggestion here and deciding it was a2 good idea ? Or did it come though other channels ?   ------------------------------  % Date: Tue, 15 Oct 2002 14:18:22 -0600 ) From: "DigiDemon" <digidemon@hotmail.com> + Subject: 7.2-1 and 7.3-1 NFS and block size , Message-ID: <aoht6102doc@enews4.newsguy.com>   Afternoon all!  C I'll go with the easy one first ;)  Just how many K or k is a block H anyways?  I would LOVE to get a size of a directory (something like du-hD from linux would rock), but I have yet to find anything like that :(  E Next question...and this one is a stumper..I'm trying to setup an NFS H server and I am having a BEAR of a time.  One is running TCPIP 5.0A (andH no documentation lucky me :( )  The other is running TCPIP 5.3.  Here is> the message I get when I try to startup the services with 5.3:  7 %TCPIP-W-INVMAP, filesystem mapping information invalid    Show map gives: ; Pathname                                Logical File System - /dra0                                   DRA0:    Show export:/ /dra0                                   WORKBOX    Show proxy: ; VMS User_name     Type      User_ID    Group_ID   Host_name 9 ROOT              OND             0           0   WORKBOX   6 Is there anywhere I can find out WHAT I'm doing wrong?  ? My last question I guess is kinda linked with the second....set I configuration map "/dra0" DRA0: doesn't get saved...as soon as I shutdown I and try and startup my map entry goes bye bye...maybe if I can finger out F how to workit right it will stay.  Thanks folks....even the docs can't seem to help me on this one!!    James    ------------------------------  % Date: Tue, 15 Oct 2002 21:46:55 -0400 2 From: rdeininger@mindspring.com (Robert Deininger)/ Subject: Re: 7.2-1 and 7.3-1 NFS and block size K Message-ID: <rdeininger-1510022146560001@1cust222.tnt3.nashua.nh.da.uu.net>   8 In article <aoht6102doc@enews4.newsguy.com>, "DigiDemon" <digidemon@hotmail.com> wrote:   >Afternoon all!  > D >I'll go with the easy one first ;)  Just how many K or k is a blockI >anyways?  I would LOVE to get a size of a directory (something like du-h E >from linux would rock), but I have yet to find anything like that :(   I A disk block in VMS is 512 bytes.  This was the memory page size on VAX.  D Alphas use 8kB (or larger) pages, and refer to the 512 byte chunk of memory as a "pagelet".    F >Next question...and this one is a stumper..I'm trying to setup an NFSI >server and I am having a BEAR of a time.  One is running TCPIP 5.0A (and I >no documentation lucky me :( )  The other is running TCPIP 5.3.  Here is ? >the message I get when I try to startup the services with 5.3:   B The TCPIP documentation is a few clicks away from the home page at    www.openvms.compaq.com 5 Also see the VMS FAQ, linked from the same home page.   D You would do well to upgrade the TCPIP 5.0A system to 5.3 as soon as> practical.  And install the ECO kit released recently for 5.3.  I It's been some time since I configured NFS.  I can't offer any reasonable F advice, except to say I figured it out from the manuals fairly easily.   ------------------------------    Date: 15 Oct 2002 15:39:48 -0600- From: koehler@encompasserve.org (Bob Koehler) . Subject: Re: Adaptec 39160 (KZPEA) & OpenVMS ?3 Message-ID: <ymo$TMerCxQz@eisner.encompasserve.org>   i In article <55f85d77.0210150450.4da6c4f9@posting.google.com>, P.Young@unsw.EDU.AU (Patrick Young) writes: K > I was not at HPETS as being on the other side of the world does not help, N > so am probably very out of date regarding the up-to-the-month state of play,K > and did believe 7.4 would be next? A google search shows information from ? > Hoff about 7.4 shipping the same time for both Alpha and VAX.   F    Public documents show Jade as 7.next (minor version number not yet @    identified) after 7.3-1 and before V.next (possibly new major    version).   ------------------------------    Date: 15 Oct 2002 13:40:05 -0600+ From: kuhrt@encompasserve.org (Marty Kuhrt) L Subject: Re: Anyone else having problems with Carepaqs and hardware support?3 Message-ID: <n$WvYTL$SHm6@eisner.encompasserve.org>   c In article <5a85bce2.0210150617.3d77d0a@posting.google.com>, svieth@wi.rr.com (Scott Vieth) writes:  > Hi:  > A > Last Tuesday night in St. Louis (HPETS2002), I attended the "HP F > Listens Executive Panel".  I stepped-up to the mic and described theE > problems that I've had recently with logging hardware service calls @ > for Compaq hardware (ES40, HSG80) that is covered by Carepaqs. > F > While describing my recent "challenges", I said "I can't be the onlyC > one having problems with Carepaqs" and then paused for a second.  A > Quite a few people in the "hall" (ballroom/auditorium/whatever) ; > applauded to indicate their agreement with my statements.  > F > So, I now ask the denizens of comp.os.vms:  Is anyone else out thereF > having problems with Carepaqs and hardware support?  The reason thatE > I'm asking is that I am gathering 'evidence' to pass along to a new  > contact that I have at HP. > @ > After the Exec Panel, I went up on the stage to talk with M.L.F > Krakauer (the HP VP for Services).  While I was waiting to talk withH > ML, a guy stepped onto the stage and introduced himself.  His name wasH > "Bob Floyd" and he's responsible for the call centers in the Americas.G >  I described my problems with Carepaqs to Bob (while ML listened) and 0 > Bob said that he would look into these issues. > D > When I returned from HPETS2002, I had a voicemail from someone whoC > works for Bob.  I spoke with this person and described my Carepaq 4 > woes.  I also have this gentleman's email address. >   ? It's Deja Vu all over again.  I was at that Listens session and ) it gave me chills to hear their response.   @ I did the same thing at the last Digital Listens and a couple of? Compaq Listens sessions over the last few years.  I didn't have = CarePaqs, but I did have 24x7, 4 hour response, contracts for 9 _all_ the hardware I purchased.  I purchased the extended 0 coverage the same time I purchased the machines.  ; Every single time I called in hardware service request they @ "couldn't find my system".  Even with service kit serial numbers? they couldn't figure it out.  I would have to brow beat them to > go out and work on the system as a per-call (i.e. we'll go out@ and work on it, but we're going to charge you the per-hour rate,: two hour minimum, and we'll get there when we get there). A Eventually they'd figure it out, but well outside of the response ( window for which I was paying a premium.  = Every time I brought this up at a Listens session I'd get the ? same response "Oh, that's terrible!  That's not right, I'll fix A it!  Give your contact information!".  Concerned blahs from upper @ management followed by more middle management blahs, followed-up by a regional management blah!  @ Next service call, same response.  Contact the person would said? they'd fix it within the organization.  Blather, rinse, repeat.   ; My "service contract" expired after three years.  I got new ? contracts with a different service provider and have been happy 
 ever since.     ? Maybe HP will be different, but my experience is thrice burned, = six times shy.  If HP would like to refund the $60k wasted on ? "extended coverage" and comp me a couple of free years hardware = maintenace to make up for Compaq and Digital's blunders, then  I'll reconsider.   ------------------------------  % Date: Wed, 16 Oct 2002 00:23:46 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> 4 Subject: Re: Are there TK50 images for VMS avalable?, Message-ID: <3DACE9C6.ACEA9090@videotron.ca>  1 > >But be warned, it [a TK50] takes AGES to boot.   L Yeah, but if you think of the TK50 as a first generation DLT drive, then the wait is not so bad anymore.   M As I recall, standalone backup wasn't THAT long, but the customer diagnostics G did take a very long time (even by all mighty Microvax II standards :-)    ------------------------------  # Date: Wed, 16 Oct 2002 00:57:21 GMT - From: "John E. Malmberg" <wb8tyw@qsl.network> 3 Subject: Re: Booting a VMS cluster node using BOOTP = Message-ID: <RP2r9.32189$m92.6204767@news1.news.adelphia.net>    Richard Brodie wrote: C  > "Christoph Gartmann" <gartmann@immunbio.mpg.de> wrote in message -  > news:ao3clm$1ib$1@n.ruf.uni-freiburg.de...   >I  >>>EWA0_protocols on the station is set to BOOTP but the machine refuses C  >>>to boot. It seems that the server is not offering a boot to the   >>>station.  >> G  >>I don't know if VMS is able to boot via BOOTP. But in order to do so G  >>the boot server must support the BOOTP protocol as well. Is this the   >>case?  G Currently booting an OpenVMS satellite node via BOOTP is not supported.   F I am not aware if there is any commitment to support it in the future,B but it seems like it would be a nice thing to have, as long as the console firmware supports it.   E  > I don't believe so. Or rather, I suppose you could persuade UCX to C  > serve the boot file to the client. However, I doubt it would fly ,  > - how would you pass the boot parameters?  H The BOOTP protocol passes the parameters for the boot loader it does notG transfer the boot image.  I have no idea how the bootp parameters would D map to what the image that is normally downloaded and started by MOP expects.  7 The name of the boot file can be one of the parameters.   F The TFTP protocol is typically used to download the bootfile and other files in a general case.   -John  wb8tyw@qsl.network Personal Opinion Only    ------------------------------    Date: 15 Oct 2002 12:08:05 -0700% From: whohe@whoever.com (DL Phillips) " Subject: Re: CTRL-Y and VMS BACKUP= Message-ID: <af0dc2ea.0210151108.7b6f94df@posting.google.com>    Phillip,  C When I restore a selected file or files from a save_set I use /LOG  B so it'll tell me that the file has been processed. Then I do a fewA CTRL-T's just to make sure that it's on to something else. Then I @ CTRL-Y and enter EXIT to close everything out. I've never had a # problem with a file when I do this.   A I, too, think that backup should be aware enough to know when it  > has processed all of the requested files, or there should be a way to tell it so.    DL Phillips  
 ------------- | Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> wrote in message news:<01KNP7VV2IIC9QWKOO@sysdev.deutsche-boerse.com>...J > I recently unpacked an RDB .RBF file (created by RMU/BACKUP) from a VMS I > BACKUP saveset.  Sizes are 2545200 and 11921805 blocks, respectively.   J > Since the file I wanted was near the beginning of the saveset, after it I > is "completely unpacked" (see below), VMS BACKUP continues to read the  J > saveset, presumably to see if there is another file by that name in the E > saveset.  From another session, I waited until DIR showed the same  I > number of USED and ALLOCATED blocks then killed the VMS BACKUP command  E > with CTRL-Y then typed EXIT.  The idea was to save resources---why  K > should BACKUP continue to read the saveset (all 12 million blocks of it)  K > when I already have the file I need?  However, the .RBF file appeared to  G > be corrupt; RMU/RESTORE said it was not a valid backup file.  Trying  K > again and letting the BACKUP command complete produced a valid RBF file;  7 > killing it with CTRL-Y again produced a corrupt file.  > F > Since "you cannot use wildcard characters denoting latest version ofJ > files (;)" with BACK/SELECT, I don't see any way to tell BACKUP to stop   > after it finds the first file. > K > Is the only alternative to let BACKUP continue to read the saveset, even  - > though my file appears to be already there?  > G > In my case, getting the file I wanted took perhaps 20 minutes and it  F > took perhaps an hour to read the whole saveset.  However, suppose I G > needed a very small file from the beginning of the saveset, which is  K > there after a few seconds, and suppose the saveset were much larger.  Do  C > I really have to read for hours after retrieving a file in a few  
 > seconds? > I > Apparently, the file whose size doesn't appear to change is changed in  J > some other way towards the end of the execution of the BACKUP command.  H > Does anyone know how and whether I can do this by hand, so that I can J > break off the BACKUP command (to save resources) and still have a valid  > SELECTed file?   ------------------------------    Date: 15 Oct 2002 12:07:57 -0700. From: spamsink2001@yahoo.com (Alan E. Feldman)" Subject: Re: CTRL-Y and VMS BACKUP= Message-ID: <b096a4ee.0210151107.2032a567@posting.google.com>   | Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> wrote in message news:<01KNP7VV2IIC9QWKOO@sysdev.deutsche-boerse.com>...J > I recently unpacked an RDB .RBF file (created by RMU/BACKUP) from a VMS I > BACKUP saveset.  Sizes are 2545200 and 11921805 blocks, respectively.   J > Since the file I wanted was near the beginning of the saveset, after it I > is "completely unpacked" (see below), VMS BACKUP continues to read the  J > saveset, presumably to see if there is another file by that name in the E > saveset.  From another session, I waited until DIR showed the same  I > number of USED and ALLOCATED blocks then killed the VMS BACKUP command  E > with CTRL-Y then typed EXIT.  The idea was to save resources---why    D Not a valid method. The size statistics you see do not mean that theA data has been completely restored. The file header information is @ restored first, which includes the size information. The data isE restored second. Use BACKUP/LOG instead to see when the file has been B fully restored. Then issue a control/y at that point, and you will have your file.   K > should BACKUP continue to read the saveset (all 12 million blocks of it)  & > when I already have the file I need?  ? In case the input file spec included commas. Such save sets may D include multiple files having the same name in different directoriesF from different disks. So, BACKUP with /SELECT is looking for the otherA possible occurences of your /SELECT file-spec (and this file-spec  can't include a disk name).   %   However, the .RBF file appeared to  G > be corrupt; RMU/RESTORE said it was not a valid backup file.  Trying  K > again and letting the BACKUP command complete produced a valid RBF file;  7 > killing it with CTRL-Y again produced a corrupt file.  > F > Since "you cannot use wildcard characters denoting latest version ofJ > files (;)" with BACK/SELECT, I don't see any way to tell BACKUP to stop   > after it finds the first file. > K > Is the only alternative to let BACKUP continue to read the saveset, even  - > though my file appears to be already there?    No.   G > In my case, getting the file I wanted took perhaps 20 minutes and it  F > took perhaps an hour to read the whole saveset.  However, suppose I G > needed a very small file from the beginning of the saveset, which is  K > there after a few seconds, and suppose the saveset were much larger.  Do  C > I really have to read for hours after retrieving a file in a few  
 > seconds?   No.   I > Apparently, the file whose size doesn't appear to change is changed in  J > some other way towards the end of the execution of the BACKUP command.  H > Does anyone know how and whether I can do this by hand, so that I can J > break off the BACKUP command (to save resources) and still have a valid  > SELECTed file?  B I would use the /LOG qualifier. When you get a message saying thatD your file has been restored, you can issue a control/y, and you will have your file.    Disclaimer: JMHO Alan E. Feldman    ------------------------------    Date: 15 Oct 2002 14:16:41 -0700. From: spamsink2001@yahoo.com (Alan E. Feldman)" Subject: Re: CTRL-Y and VMS BACKUP= Message-ID: <b096a4ee.0210151316.35906a2c@posting.google.com>   | Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> wrote in message news:<01KNP7VV2IIC9QWKOO@sysdev.deutsche-boerse.com>...J > I recently unpacked an RDB .RBF file (created by RMU/BACKUP) from a VMS I > BACKUP saveset.  Sizes are 2545200 and 11921805 blocks, respectively.   J > Since the file I wanted was near the beginning of the saveset, after it I > is "completely unpacked" (see below), VMS BACKUP continues to read the  J > saveset, presumably to see if there is another file by that name in the E > saveset.  From another session, I waited until DIR showed the same  I > number of USED and ALLOCATED blocks then killed the VMS BACKUP command  E > with CTRL-Y then typed EXIT.  The idea was to save resources---why  K > should BACKUP continue to read the saveset (all 12 million blocks of it)  K > when I already have the file I need?  However, the .RBF file appeared to   [...] I > Apparently, the file whose size doesn't appear to change is changed in  J > some other way towards the end of the execution of the BACKUP command.  H > Does anyone know how and whether I can do this by hand, so that I can J > break off the BACKUP command (to save resources) and still have a valid  > SELECTed file?    C Sorry, I have to correct myself again. In my first response to this E post, I said that the file header info is restored first, followed by D the data. However, I since found out that I was wrong (apologies forA jumping the gun). Apparently, only the size, and just a few other E details about the file are restored first. Then the data is restored. F Then the remaining details, including *original* dates, record format, and others, are restored.   C So if you press control/Y during a restore and you happen interrupt B the restoration of a particular file, that file will likely be not? only incomplete, but also have incorrect header info, including 0 incorrect dates. So be careful when you do that.  D Nevertheless, my suggestion of using BACKUP/LOG is the answer to the; poster's question. Any file that is reported as having been 5 successfully created was, well, successfully created.    Disclaimer: JMHO Alan E. Feldman    ------------------------------    Date: 15 Oct 2002 14:59:07 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)  Subject: Re: DLM questions= Message-ID: <cf15391e.0210151359.5115eb2f@posting.google.com>   f kuhrt@encompasserve.org (Marty Kuhrt) wrote in message news:<dMv38OplTuAu@eisner.encompasserve.org>...@ > I've been working on a minor performance problem on a two nodeC > SCSI cluster.  Seems that accessing a specific file from one node : > is drastically slower than the other during high traffic? > conditions.  I did a MON LOCK/ALL/REC during the time when we > > could reproduce the problem, and have started looking at the
 > results.  C $MONITOR DLOCK might also be of interest, since you're in a cluster   and multiple nodes are involved.  D As an initial guess, I'd say the file of interest has its associatedD RMS resource/lock tree mastered on the node with the better response time.   F You didn't say what type of cluster interconnect is in use between theF nodes.  It might be interesting to run Roy Davis' LOCKTIME_1/3 programC pair and measure the average lock request latency between nodes, to : see if it's out of the ordinary.  You can grab a copy fromC http://encompasserve.org/~parris/locktime_1.mar and *_3.mar.  Raise E ENQLM on some account to over 5K, then run LOCKTIME_3 on one node (it D becomes the master for some test lock trees), then run LOCKTIME_1 onD the other (to measure lock latencies), then press Return in responseC to the LOCKTIME_3 prompt on the first node to allow that program to C end.  Look at the lock conversion time as an estimate of round-trip C lock latency (when there's no contention).  If you report back here E with the interconnect type and latency, we can get an idea of whether E or not it's normal.  LAN interconnects will tend to be in the 200-300  microsecond range.  5 > I'm familiar with how LOCKDIRWT and PE1 will affect 8 > lock traffic, but there are some questions that remain
 > unanswered.   @ If the slowness symptoms moved back and forth between nodes, I'dC suspect the lock tree involved was being remastered.  But from your = description, it doesn't sound like that's happening.  Are you C concerned that you're seeing lock mastership thrashing and that the D extra work due to this is slowing down lock requests for the file of	 interest?   D You didn't say what the values of LOCKDIRWT were on your nodes.  AreF you concerned that an imbalance in LOCKDIRWT is keeping the mastershipE for the resource tree for this file on one node when it might best be  migrated to the other?  B If you'd like to see the locking activity rates on the active lockD trees, including the one for your file of interest, as well as whichB node they're mastered on, you could use my Lock Activity tool (seeE http://encompasserve.org/~parris/lckact.bck for the latest version).  : If you had OpenVMS version 7.2-2 or greater, you could use    SDA> LCK SHOW ACTIVE   B > My first question is where should I start looking to address theB > fields in the output titled "ENQs Forced To Wait Rate" and "ENQs@ > Not Queued Rate"?  I did a Google search through COV and found= > some info, but nothing that said "High values here mean the 6 > FRAMMINJAM SYSGEN parameter is too low", or similar.  A "ENQs Forced To Wait" are lock requests that must wait, typically B because another process is presently holding a lock in a mode thatA precludes this process's proposed mode of access to the resource.   D "ENQs Not Queued" are cases where the programmer said that if a lockF could not be immediately granted, to return to the program to allow itA to take some sort of action, instead of just waiting for the lock ! request to eventually be granted.   C What you really want to know in either case is what resource you're 8 contending for, and MONITOR doesn't tell you that.  (AndA unfortunately, in the "Not Queued" case, the program doesn't wait C around, so you can't easily see what it was encountering contention  on.)  Using    SDA> SHOW RESOURCE/CONTENTION A may help, although it's a bit much to wade through.  If locks are D blocked for more than about 15 seconds, DECamds/Availability ManagerC should detect them in its Lock Contention data collection.  Failing / that, my Lock Queue tool can be very handy (see F http://encompasserve.org/~parris/lckque.bck for the latest version) inE pointing out lock queues.  It provides a snapshot as of an instant in E time, so you'd probably want to run it every minute or so during busy  times to gather some samples.   F You should probably also check to see if $SHOW CLUSTER/CONTINUOUS withB ADD CONNECTIONS and ADD CR_WAITS shows a lot of credit waits.  The; lock manager uses the VMS$VAXcluster SYSAP, and raising the @ CLUSTER_CREDITS parameter might help in that case.  Insufficient> credits might cause lock requests to be unnecessarily delayed.  B > My second question has to do with PE1.  This determines the size@ > of the largest lock tree that can be moved.  How do you figure- > out the size of the lock trees on a system?   D You can use SDA> SHOW RESOURCE and look at the lock counts, for eachC level of a resource tree, but that's a lot of work.  People who are E experiencing excessive remastering often pick a "large" value for PE1 ? and gradually lower it (it's dynamic) until the pauses or other - negative symptoms they are observing go away.   D If you were seeing excessive remastering activity, you'd likely spotB processes in RWCAP state.  You'd also see counters incrementing by, large amounts fairly often when you look at:#    SDA> EXAMINE PMS$GL_RM_RBLD_SENT #    SDA> EXAMINE PMS$GL_RM_RBLD_RCVD " If you were running 7.2 or higher,    SDA> SHOW LOCK/SUMMARY F would be helpful, or if 7.3 or higher, as someone already pointed out, $MONITOR RLOCK would help here.   F It's also quite possible that something other than locking is involvedC in the difference in response times.  For example, if one node were E accessing the data via an MSCP-served path, that would be slower.  Or F the file were an RMS file, global buffers were in place, and there wasE a lot more I/O activity on one node, processes on that node might get E higher cache hit rates.  You could use $SET FILE/STATISTICS to enable C RMS statistics-gathering, and $MONITOR RMS/FILE=xxx/ITEM=CACHING to B look at cache hit rates and see how they differ between the nodes.   ------------------------------  % Date: Tue, 15 Oct 2002 22:52:07 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> " Subject: Re: Hobbyist kit Contents, Message-ID: <3DACD451.B1AF760B@videotron.ca>   John Smith wrote: J > What's wrong with Montgar placing images on a DVD that costs $5 to make.  A Hobbyists tend to use older machines which don't have DVD drives.    ------------------------------  % Date: Tue, 15 Oct 2002 21:21:17 -0400 2 From: rdeininger@mindspring.com (Robert Deininger)- Subject: Re: HoPping along with two left feet K Message-ID: <rdeininger-1510022121180001@1cust222.tnt3.nashua.nh.da.uu.net>o  D In article <3DA5984F.8010203@caltech.edu>, mathog@caltech.edu wrote:  = >A nice lady called me yesterday to inquire if I was going tol> >be attending the Virtual Memory System event on October 30th. ...L >A; >The folks in VMS engineering might want to investigate whol: >HP has making these calls and why they were given such an >odd script to read from.   J You could probably save somebody a lot of time if you revealed the name of the lady who called you.   ------------------------------  % Date: Tue, 15 Oct 2002 22:00:27 -0400S- From: JF Mezei <jfmezei.spamnot@videotron.ca>Y8 Subject: Re: How hard is this for HP to do?  Clustering?, Message-ID: <3DACC838.170F1DE5@videotron.ca>   Neil Rieck wrote:fL > Every day it seems I see or hear evidence that HP (an engineering company)M > is better home for VMS than Compaq (a Pee Cee company). Maybe it's just me,VF > but I still see Ken Olsen in the same light as Bill Hewlett and Dave# > Packard; a bunch of cool dudes...t  B The company Hewlette and PAckard founded is now called "Agilent".   K Reminds me a bit of the "Hudson Bay Company" in Canada. It claims to be thetK oldest corporation in canada. Problem is that it is legalese only since the M corporation operates no longer operates any of the stuff it was doing when ite@ was founded (fur trading and remote outposts to sell gear/food).  N I am more confused than ever on what Carly really intends to do with "her" HP.I  Prior to the merger being consumed, she was putting a lot of emphasis onnM industry standard and wintel junk. (as well as printers and  "imaging"). Thate> picture of HP has absolutely nothing to do with the "real" HP.  N But now,  I haven't heard the "Wintel will eviscerate" type of statement for aX while, and there have been reports of HP taking the real enteprise stuff more seriously.  I On the other hand, a recent statement said that HP was no losing money onrN enterprise systems and would focus much of the cuts on those, and wasn't doingL much profit on the consulting part (isn't that what gave Digital/Compaq much+ of the profits in the "services" division ?     L So, what exactly is HP ? Are they going to downsize themselves to focus onlyK on printers and imaging ? That seems to be the only consistent product linem that HP wants to focus on.   ------------------------------  % Date: Tue, 15 Oct 2002 20:04:17 +0200e6 From: Arne =?ISO-8859-1?Q?Vajh=F8j?= <arne@vajhoej.dk>@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!% Message-ID: <3DAC58A1.200@vajhoej.dk>n   Fred Kleinsorge wrote:  K > Tell you what Andy.  Lets have a vote.  I'll abide by it.  If you fail toa9 > get a majority NO vote, then we'll each just fade away.v > H > Everyone - please replay with your votes.  I'll abide by the majority. > K > Remember a YES vote means "Please go away".  A NO vote means "Please stayx > and keep writing". >  >   > Fred is a moron:    YES  /  NO > Andy is a moron:    YES /  NOf   Fred is a moron:    NO Andy is a moron:    YES    Arne  5 PS: But there are not a snowball in hells chance that #      Andrew will accept the result.o   ------------------------------  % Date: Tue, 15 Oct 2002 12:44:58 -0600e$ From: Lee Y T Mah <lytmah@cha.ab.ca>@ Subject: RE: HP tells VMS users "Relax, you are sitting pretty"!) Message-ID: <3DAC6229.93D6C343@cha.ab.ca>A   Fred is a moron:    NO Andy is a moron:    YESg   -- Leea  ; Lee Y T Mah                        Capital Health Authoritys? Email: lytmah@cha.ab.ca            Information Systems, RAH CSC-4 Phone:  (780) 477-4725, 477-4233   10240 Kingsway NW? Fax:      (780) 491-5119, 491-5619    Edmonton, AB, CAN  T5H3V9:   ------------------------------  % Date: Tue, 15 Oct 2002 21:16:15 +0200C9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>o@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!' Message-ID: <3DAC697F.75C4CFAC@aaa.com>   < Well, anyone actualy answering Y/N on those questions are...   Jan-Erik Sderholm   >   > Fred is a moron:    YES  /  NO > Andy is a moron:    YES /  NOk >    ------------------------------  % Date: Tue, 15 Oct 2002 12:22:23 -0700u' From: David Mathog <mathog@caltech.edu>s@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!* Message-ID: <3DAC6AEF.1010508@caltech.edu>   Fred Kleinsorge wrote:   >   > Fred is a moron:    YES  /  NO!  > Andrew is a moron:  YES  /  NOe   This thread is moronic:  YES   -- e David Mathog mathog@caltech.edu> Manager, Sequence Analysis Facility, Biology Division, Caltech   ------------------------------  # Date: Tue, 15 Oct 2002 20:33:47 GMTs5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com>,@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!2 Message-ID: <LY_q9.42$Gz1.861277@news.cpqcorp.net>  B Awww.  Come on Dave, it's better than the grab bag of domestic and5 international political garbage that was going on ;-)    _Fred-  ? David Mathog wrote in message <3DAC6AEF.1010508@caltech.edu>...r >Fred Kleinsorge wrote:h >o >>! >> Fred is a moron:    YES  /  NOi" > > Andrew is a moron:  YES  /  NO >3 >This thread is moronic:  YES  >3 >--L
 >David Mathogi >mathog@caltech.edu ? >Manager, Sequence Analysis Facility, Biology Division, CaltechM >R   ------------------------------    Date: 15 Oct 2002 15:42:16 -0600- From: koehler@encompasserve.org (Bob Koehler)w@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!3 Message-ID: <Z0CLFGDahDuR@eisner.encompasserve.org>n  j In article <KoYq9.27$qn1.516127@news.cpqcorp.net>, "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> writes:   > Andy is a moron:    YESr   ------------------------------    Date: 15 Oct 2002 16:06:52 -06003 From: mcdermott@encompasserve.org (C. M. McDermott) @ Subject: RE: HP tells VMS users "Relax, you are sitting pretty"!3 Message-ID: <in2lk20+BdxV@eisner.encompasserve.org>:  i In article <1E4B06029E11D211B47C0000F8207F4D01B9F448@ESKC2>, "Stewart, Bill" <wjs-corp@Kaman.com> writes:o   <AOL>h   > Fred is a moron:    NO > Andy is a moron:    YES    </AOL>   ------------------------------  % Date: Tue, 15 Oct 2002 18:40:56 -0400t* From: "Bill Todd" <billtodd@metrocast.net>@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!6 Message-ID: <FEqdnWAZHPT7BDGgXTWc3A@News.GigaNews.Com>  @ "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote in message, news:KoYq9.27$qn1.516127@news.cpqcorp.net...   ...:    > Fred is a moron:    YES  /  NO > Andy is a moron:    YES /  NO2  @ I'm still reserving final judgement in both cases.  But based onI currently-available information, I suspect that whatever the answer is it  will be the same for both.   - bill   ------------------------------    Date: 15 Oct 2002 16:50:30 -0700( From: bob@instantwhip.com (Bob Ceculski)@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!< Message-ID: <d7791aa1.0210151550.4ae4d09@posting.google.com>  o "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote in message news:<KoYq9.27$qn1.516127@news.cpqcorp.net>...0K > Tell you what Andy.  Lets have a vote.  I'll abide by it.  If you fail toR9 > get a majority NO vote, then we'll each just fade away.- > H > Everyone - please replay with your votes.  I'll abide by the majority. > K > Remember a YES vote means "Please go away".  A NO vote means "Please stay  > and keep writing". >  >   > Fred is a moron:    YES  /  NO > Andy is a moron:    YES /  NOR >   
 FRED    NO ANDY    YES   ? but we need andy to keep posting because he contributes so much0< to vms marketing ... every time he posts, vms sales grow ...   ------------------------------  % Date: Tue, 15 Oct 2002 17:26:24 -0700m$ From: Shane Smith <ssmith@icius.com>@ Subject: RE: HP tells VMS users "Relax, you are sitting pretty"!0 Message-ID: <01C27470.01B03700@sulfer.icius.com>  A All the YES/NO posts I've seen so far are I think underestimating0F Andrew. Is he anti-VMS? Yes. Is he usually spreading FUD? Yes. Does heE let his emotions get in the way of good sense? Oh yes. Do I like him?eE Hell no. But I don't think he's as dumb as some people think. There'st@ often some very good points buried among his other (ahem) stuff.  G Any good military manual will advise you not to underestimate an enemy.t   Shanea   -----Original Message-----6 From: bob@instantwhip.com [mailto:bob@instantwhip.com]' Sent: Tuesday, October 15, 2002 4:51 PMf To: Info-VAX@Mvb.Saic.ComP@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!    @ "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote in message. news:<KoYq9.27$qn1.516127@news.cpqcorp.net>...K > Tell you what Andy.  Lets have a vote.  I'll abide by it.  If you fail to 9 > get a majority NO vote, then we'll each just fade away.  > H > Everyone - please replay with your votes.  I'll abide by the majority. > K > Remember a YES vote means "Please go away".  A NO vote means "Please staym > and keep writing". >  >   > Fred is a moron:    YES  /  NO > Andy is a moron:    YES /  NOt >   
 FRED    NO ANDY    YESs  ? but we need andy to keep posting because he contributes so muchr< to vms marketing ... every time he posts, vms sales grow ...   ------------------------------  % Date: Tue, 15 Oct 2002 20:59:14 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>k@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!' Message-ID: <3DACC7F2.9F7530E8@fsi.net>e   Fred Kleinsorge wrote: > [snip]  > Fred is a moron:    YES  /  NO > Andy is a moron:    YES /  NO2   Fred: No	 Andy: Yeso  G Actually, IMO, it would be more correct to say that Andy is a trolling,dH argumentative putz, but that wasn't one of the options. So, "moron" will have to suffice.   -- t David J. Dachteral dba DJE Systemsw http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/a   ------------------------------  % Date: Tue, 15 Oct 2002 21:04:05 -0500o1 From: "David J. Dachtera" <djesys.nospam@fsi.net> @ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!' Message-ID: <3DACC915.326A4B48@fsi.net>5   Shane Smith wrote: > C > All the YES/NO posts I've seen so far are I think underestimatingrH > Andrew. Is he anti-VMS? Yes. Is he usually spreading FUD? Yes. Does heG > let his emotions get in the way of good sense? Oh yes. Do I like him?nG > Hell no. But I don't think he's as dumb as some people think. There'srB > often some very good points buried among his other (ahem) stuff. > I > Any good military manual will advise you not to underestimate an enemy.o  G Agreed. As I posted earlier, "moron" is the wrong word. Fred is usuallym@ helpful, but he gets set in a track sometimes. Andrew is usuallyH longwinded and thin on content, and tries to turn anything he can to his own advantage.  @ I'd have to say that IMO, Andrew is the cuplrit in as much as heG provokes fairly consistently, regardless of who is the sparring partner-	 du jour.     -- @ David J. DachteraB dba DJE Systemsj http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/n   ------------------------------  % Date: Wed, 16 Oct 2002 00:19:40 -0400U- From: JF Mezei <jfmezei.spamnot@videotron.ca>E@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!, Message-ID: <3DACE8D0.D9219669@videotron.ca>   Fred Kleinsorge wrote: > Fred is a moron:    NO > Andy is a moron:    NO  I I don't mind the catfights between you two. Since Fred is a VMS god (aka:AL engineer), I must listen to him. But I don't mind listening to what Andy hasM to say, sometimes he brings a different viewpoint that is valid. Sometimes it ? is just propaganda, but such propaganda comes from all vendors.   H And knowing what the competitor's propaganda is is valuable information.   ------------------------------    Date: 15 Oct 2002 15:34:40 -0600- From: koehler@encompasserve.org (Bob Koehler) ) Subject: Re: Java performance on open VMSo3 Message-ID: <q5xX1lWNtkAT@eisner.encompasserve.org>t  u In article <7339437.0210150357.335e0605@posting.google.com>, david.rabjohns@logistics.nhs.uk (David Rabjohns) writes:p > Hi > E > We have a java multithreaded server program that listens for simplegC > commands and returns data from RMS files by spawning a process (a B > command procedure) that runs a (vax basic) exe and capturing itsB > output. This works fine at the moment but is suffering from poor > performance. > H > I have considered the follwing possibilities but am struggling to find* > the info I need to put it into practice. > C > 1. "Select" the required data directly from the java class, but I H > can't seem to find any information on retrieving data from indexed RMSE > files from within Java if at all this is possible. I have looked attE > j2VMS which may fit the bill but I'm not sure I fully understand itx > yet. Any other ideas ?  I    The Java I/O library is built on top of the C/C++ RTL which uses RMS. aH    You can read RMS files, but you cannot read by key since C, C++, and E    Java are brain dead to that.  Sequential reads to a keyed indexed tJ    file should be returned in order as if the data had been sorted by key.  @    Get a good look at the size of your file?  Often times I findA    a great deal of concern over a several megabyte file.  If it'so>    that small read it all in and access it directly in memory.  A > 2. Use JNI to call the vax basic "methods" directly rather than D > spawning another process that executes an image etc.... (We do notE > have a C license on the cluster) But I can not find any examples ofhF > doing this with vax basic. Has anyone out there integrated vax basic) > procedures with Java ? Is it possible ?   G    Someone once posted the few lines of code needed to do a Java nativepC    class accessing RMS keyed-indexed files by key.  Look in Google.s  B > If anyone has any other JAVA/VMS performance tips or can suggestA > things I may have overlooked - they will be greatly appreciatedl  D    Don't use Java when performance is important.  I've actually beenC    suprized how fast it runs, but it won't get anywhere near nativen    code.  H    gnu is working on a Java compiler that will optionally produce nativeE    code.  That's built on top of gcc so it may be able to write AlphabF    VMS object files if they ever get it done.  Can you wait that long?   ------------------------------  % Date: Tue, 15 Oct 2002 23:37:06 +0200r6 From: Arne =?ISO-8859-1?Q?Vajh=F8j?= <arne@vajhoej.dk>) Subject: Re: Java performance on open VMSC) Message-ID: <3DAC8A82.1060802@vajhoej.dk>o   Bob Koehler wrote:  w > In article <7339437.0210150357.335e0605@posting.google.com>, david.rabjohns@logistics.nhs.uk (David Rabjohns) writes:aF >    Don't use Java when performance is important.  I've actually beenE >    suprized how fast it runs, but it won't get anywhere near native 
 >    code.    / If the JVM can get enough physical memory, thene it is reasonable fast.  9 The worst performing Java apps is not computing intensiveh& apps as many will expect but GUI apps.    dJ >    gnu is working on a Java compiler that will optionally produce nativeG >    code.  That's built on top of gcc so it may be able to write AlphaeH >    VMS object files if they ever get it done.  Can you wait that long?    + GCJ has been available for quite some time.a  1 It seems to work, but the EXE files are very big.l   But no VMS support.C   Arne   ------------------------------  # Date: Tue, 15 Oct 2002 18:30:43 GMT 4 From: Tim Llewellyn <tim.llewellyn@blueyonder.co.uk> Subject: Re: Life after VMS?0 Message-ID: <3DAC5C90.C06F7994@blueyonder.co.uk>   Phillip Helbig wrote:b  tH > to a different---and this is the important bit---NON-HP PLATFORM.  Ask  v7 not easy if they already have a big commitment to hpux.    -- e tim.llewellyn@blueyonder.co.uk m  H * PLEASE NOTE tim.llewellyn@cableinet.co.uk address is NO LONGER VALID *   ------------------------------  + Date: Tue, 15 Oct 2002 20:50:20 +0100 (MET)t9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>m Subject: Re: Life after VMS?; Message-ID: <01KNPH1QBLDY9QWKOO@sysdev.deutsche-boerse.com>   J > > to a different---and this is the important bit---NON-HP PLATFORM.  Ask >  b9 > not easy if they already have a big commitment to hpux.   H The original poster mentioned a firm concerned about HP's commitment to @ VMS and whether they should migrate VMS applications to another I platform.  Of course, there is nothing special about hpux which makes it iF suitable as a migration target for VMS applications (aside, possibly, H from the fact that hpux might get some clustering technology from tru64 H which tru64 got from VMS).  Even if the company has a strong commitment H to hpux (presumably for other applications), if they've had two vendors G (DEC and HP) up to now, they probably won't wince at having two in the uH future.  If anything, if they have doubts about HP's commitment to VMS, = they might have doubts about its commitment to tru64 as well.    ------------------------------  % Date: Tue, 15 Oct 2002 21:46:13 +0200 - From: Didier Morandi <Didier.Morandi@Free.fr>. Subject: Re: Life after VMS?' Message-ID: <3DAC7084.1A994759@Free.fr>P  
 VAXVMS wrote:c >  >      Didier: > ) >      Have you contacted the EMEA folks?s. >      I've always found them to most helpful.  N As this is a public forum, I cannot comment this question, but let me just sayO that DEC folks within CPQ was not that easy to reach and get involved. Now, HPQ-O folks who have heard about VMS are really today in a shrinking handkerchief :-)oN and there is no more VMS advisory support over here. More offline if you wish.  N >      If nobody gives you the large-caliber assistance that you need, there'sM >      a very nice lady who has the same first and last initials who might beI( >      able to step in and do something.  M Yeah, I know, but I try not to bother her too much. What could be a good deal M for VMS Customers in Europe (and new to come) would be for HP to hire me as at VMS EMEA VP :-)i  O I'll escalate to Corporate when the Customers start saying (as Michelin told mem[ a few months ago) that their CPQ office does not call them back when they leave messages....   D.   ------------------------------  % Date: Tue, 15 Oct 2002 21:52:36 +02000- From: Didier Morandi <Didier.Morandi@Free.fr>  Subject: Re: Life after VMS?' Message-ID: <3DAC7203.6CB96765@Free.fr>s   "Main, Kerry" wrote: >  > A few suggestions -i >  ../..d  K These are marketing stuff, so not trustable. A 1/2 million $ TV ad would behM trustable. A few years ago, when you talked to the Field in France about VMS,uK answer was "pfff" with a shoulders shake. Today, answers are "what is it?".h   > Email: kerryDOTmain@hpDOTcom >   (remove the DOT's)  C %SMTP-F-SYNTAX, Error parsing kerrymain@hpcom, invalid email syntaxo :-)f   D.   ------------------------------  % Date: Tue, 15 Oct 2002 21:57:00 +0200o" From: "Hans Vlems" <hvlems@iae.nl> Subject: Re: Life after VMS?5 Message-ID: <aohrv8$ml81a$1@ID-143435.news.dfncis.de>l   Didier,d  J there's no argument whatsoever that you can think of and that will be ableL to counter a sense of fear at upper management levels. Fear of selecting theD "wrong" product because the vendor may fold or the pruct will becomeA obsolete, or whatever fear flavor they're subscibed to that week.vH I think it was Christoph who mentioned to take them to another VMS site.G Preferably a site comparable in size (employees, turnover) in a similaroI market segment, but preferably not a competitor. It would be very nice ifv> that other site has some kind of "quality" associated with it.< Example: Nedcar builds Volvo's (S/V 40 series) and uses VMS.   Hans  < "Didier Morandi" <Didier.Morandi@Free.fr> schreef in bericht! news:3DABAE5C.78F2C057@Free.fr...fK > I may be soon in the process of visiting a laaarge aircrafts manufacturereH > company who is wondering what they should do if VMS is dead (again...) >oL > After having thought about this (known) issue, I think that I will propose three choices: >aL > 1. the "Wait&See" solution. Do nothing. As long as you have support, spare partsED > and no dysfunctioning (sp?), there is no hurry to change anything. >7L > 2. the "Start-That-Big-Project" solution. Hire a high level consultant for two J > or three years (me, if possible) and he will tell you where the problems couldeK > raise up going to this or that platform, then he will prepare a good cook9 toK > migrate/rewrite/adapt/translate/etc the applications (as I know that theya haveJ > *plenty* of real time stuff (and maybe also some PDPs around) it's gonna be1 > funny. But PDPs/RSX11M/S is not VMS, is it? :-)  >r > 3. theH "Please-Trust-That-New-HP-Company-Folks-They-Are-Not-That-Bad-After-All"D > (ie Rich M. and Co) solution, where I will tell the Customer (with
 documents andlI > a little help from my ZK friends): "Please stop wondering about "and ifR VMS is4 > dead" and continue the good job, we are with you". >l > Comments?A >P > D. > --4 >   ------------------------------------------------4 > MORANDI Consultants  http://Didier.Morandi.Free.fr2 >   19 chemin de la Butte, 31400 Toulouse, France.4 > Tel.: +33 (0)6 7983 6418 - Fax: +33 (0)5 6154 19284 > OpenVMS, APPLE, Computer Security, Migration plans4 > --------------------------------------------------   ------------------------------    Date: 15 Oct 2002 15:26:54 -0600- From: koehler@encompasserve.org (Bob Koehler)  Subject: Re: Life after VMS?3 Message-ID: <AJtiH1NsOSMT@eisner.encompasserve.org>I  W In article <3DABAE5C.78F2C057@Free.fr>, Didier Morandi <Didier.Morandi@Free.fr> writes:tK > I may be soon in the process of visiting a laaarge aircrafts manufacturer H > company who is wondering what they should do if VMS is dead (again...) > [ > After having thought about this (known) issue, I think that I will propose three choices:  > R > 1. the "Wait&See" solution. Do nothing. As long as you have support, spare partsD > and no dysfunctioning (sp?), there is no hurry to change anything.  G    1.  Wait and plan.  Wait to see if HP's IPF plans pan out and plan a E       port to IPF if they do.  In the meantime their existing systemsgE       won't die and they can evaluate the portability of their stuff.tG       Very important: does thier stuff require real-time high interruptlC       rate interfaces?  High security?  High uptime?  Point them ath;       that HP slide set on VMS which just hit the web site.   P > 2. the "Start-That-Big-Project" solution. Hire a high level consultant for twoP > or three years (me, if possible) and he will tell you where the problems couldN > raise up going to this or that platform, then he will prepare a good cook toP > migrate/rewrite/adapt/translate/etc the applications (as I know that they haveM > *plenty* of real time stuff (and maybe also some PDPs around) it's gonna beo1 > funny. But PDPs/RSX11M/S is not VMS, is it? :-)   B       Well, I guess that's what I just said.  It's the size of theG       project that dictates whether they need to start it now, how manya       consultants, and such.  Q > 3. the "Please-Trust-That-New-HP-Company-Folks-They-Are-Not-That-Bad-After-All" R > (ie Rich M. and Co) solution, where I will tell the Customer (with documents andP > a little help from my ZK friends): "Please stop wondering about "and if VMS is4 > dead" and continue the good job, we are with you".  D    Never put your trust in some other company.  Not HP, not IBM, not(    Sun, not Gartner, ...  Wait and plan.   ------------------------------  % Date: Tue, 15 Oct 2002 22:05:18 +0200b- From: Didier Morandi <Didier.Morandi@Free.fr>n Subject: Re: Life after VMS?' Message-ID: <3DAC74FD.B434760F@Free.fr>e   Phillip Helbig wrote:B ../.3 > if they have doubts about HP's commitment to VMS,t? > they might have doubts about its commitment to tru64 as well.c  N No, because Unix is recognized in Europe as an unavoidable standard even if it is not.c  O You know, I just spent two full days to fix a friend's bloody PC/W98 hit by theiP BugBear virus. I had to do this and that and reboot more than 100 times until itJ starts to run again and allow Norton Antivirus to get installed (the virusM changed the name of a required library to prevent installation) and set up ansP Internet connection to update the NAV definition and finally give something elseM than a 640x480 16 colors image (and I still do not have the sound because theBO card driver was lost and the vendor CD is lost too and the driver I found gives O a conflict with a Windows driver, and I also miss the second HD because it's anBN IDE CAVIAR 21600 and the driver disquette is lost too and it is no more on the$ Western Digital site and and and...)  L As long as the Field considers that the office platform IS Windows, there is still no truth on this planet.
 (no flame)   D. made with a Mac    ------------------------------    Date: 15 Oct 2002 15:21:21 -0600- From: koehler@encompasserve.org (Bob Koehler)n Subject: Re: Life after VMS?3 Message-ID: <ByRAkbLEYOnj@eisner.encompasserve.org>   W In article <3DABAE5C.78F2C057@Free.fr>, Didier Morandi <Didier.Morandi@Free.fr> writes:gK > I may be soon in the process of visiting a laaarge aircrafts manufacturerdH > company who is wondering what they should do if VMS is dead (again...)  E    Tell them to waste their money someplace else.  Even if HP abandony2    VMS their systems will run longer than an A300.    n   ------------------------------    Date: 15 Oct 2002 16:27:57 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)  Subject: Re: Life after VMS?3 Message-ID: <SOllDmkrlywX@eisner.encompasserve.org>c  W In article <3DAC7203.6CB96765@Free.fr>, Didier Morandi <Didier.Morandi@Free.fr> writes:- > "Main, Kerry" wrote: >>   >> A few suggestions - >> r > ../..p > M > These are marketing stuff, so not trustable. A 1/2 million $ TV ad would be1 > trustable.  @ I would trust the DII-COE committment to the US Government more.  C In France, you may not care about the US Government, but in the US, 	 HP cares.    ------------------------------  # Date: Tue, 15 Oct 2002 23:10:21 GMTe# From: "John Smith" <a@nonymous.com>m Subject: Re: Life after VMS?B Message-ID: <xf1r9.11$mxk1.4@news04.bloor.is.net.cable.rogers.com>  : "Didier Morandi" <Didier.Morandi@Free.fr> wrote in message! news:3DAC7084.1A994759@Free.fr...t > VAXVMS wrote:  > >N > >      Didier: > >e+ > >      Have you contacted the EMEA folks? 0 > >      I've always found them to most helpful. >IL > As this is a public forum, I cannot comment this question, but let me just saytH > that DEC folks within CPQ was not that easy to reach and get involved. Now, HPQ@ > folks who have heard about VMS are really today in a shrinking handkerchief :-)J > and there is no more VMS advisory support over here. More offline if you wish., > H > >      If nobody gives you the large-caliber assistance that you need, there'soL > >      a very nice lady who has the same first and last initials who might be* > >      able to step in and do something. >hJ > Yeah, I know, but I try not to bother her too much. What could be a good dealJ > for VMS Customers in Europe (and new to come) would be for HP to hire me as a > VMS EMEA VP :-)h    L Great idea. Be prepared to sit in your office all day because you won't haveH a budget for calling on prospective *new* customers, nor will you have aG telephone to call them with, nor will you have a budget for demo units,r etc...  H But think of the positives -  it won't be stressful because you won't beK competing against Sun and IBM for business. There won't be any deadlines tocJ meet for RFP's or important presentations to customers. You could probablyK sit on the beach in Menton all day long and still meet your sales quota. HPo< probably wouldn't even know that you were out of the office.   ------------------------------  # Date: Tue, 15 Oct 2002 23:15:05 GMTg# From: "John Smith" <a@nonymous.com>h Subject: Re: Life after VMS?C Message-ID: <Zj1r9.21$mxk1.12@news04.bloor.is.net.cable.rogers.com>s  : "Bob Koehler" <koehler@encompasserve.org> wrote in message- news:AJtiH1NsOSMT@eisner.encompasserve.org...p > F >    Never put your trust in some other company.  Not HP, not IBM, not* >    Sun, not Gartner, ...  Wait and plan.    K The corollary to this is never trust an employer or a vendor. Never believegL what a vendor says unless it is in the contract with financial penalties forG non-compliance. Keep your resume up-to-date and 3-6 months salary in anr emergency fund.    ------------------------------  % Date: Tue, 15 Oct 2002 16:43:39 -0700u$ From: Shane Smith <ssmith@icius.com> Subject: RE: Life after VMS?0 Message-ID: <01C2746A.0B8C61A0@sulfer.icius.com>  # Oh Hell Yeah; words of true wisdom.t  F Speaking of which, I have a friend in the Panorama City (maybe 40 minsH north of LA) area who recently found out how important the last sentenceH is. Anybody know of a company who's looking for experienced, trustworthy operations staff?    Shane    -----Original Message-----( From: John Smith [mailto:a@nonymous.com]' Sent: Tuesday, October 15, 2002 4:15 PMo To: Info-VAX@Mvb.Saic.Com  Subject: Re: Life after VMS?      : "Bob Koehler" <koehler@encompasserve.org> wrote in message- news:AJtiH1NsOSMT@eisner.encompasserve.org...t >lF >    Never put your trust in some other company.  Not HP, not IBM, not* >    Sun, not Gartner, ...  Wait and plan.    C The corollary to this is never trust an employer or a vendor. Nevern believe H what a vendor says unless it is in the contract with financial penalties forsG non-compliance. Keep your resume up-to-date and 3-6 months salary in an  emergency fund.e   ------------------------------  # Date: Wed, 16 Oct 2002 01:35:12 GMTh1 From: LESLIE@JRLVAX.HOUSTON.RR.COM (Jerry Leslie)l Subject: Re: Life after VMS?< Message-ID: <kn3r9.140433$8o3.4062582@twister.austin.rr.com>  " John Smith (a@nonymous.com) wrote: : < : "Bob Koehler" <koehler@encompasserve.org> wrote in message/ : news:AJtiH1NsOSMT@eisner.encompasserve.org...t : >iH : >    Never put your trust in some other company.  Not HP, not IBM, not, : >    Sun, not Gartner, ...  Wait and plan. :  : F : The corollary to this is never trust an employer or a vendor. Never I : believe what a vendor says unless it is in the contract with financial nK : penalties for non-compliance. Keep your resume up-to-date and 3-6 months A : salary in an emergency fund. :   F That should be 12-24 months salary, since this recession is worse thanH the previous four recessions according to the Economic Policy Institute:  5    http://www.epinet.org/press/releases/ui071602.htmleD    NEW REPORT SHOWS LONG-TERM UNEMPLOYMENT MORE SEVERE THAN IN PAST !    RECESSIONS | EPI Press Release   G   "WASHINGTON, DC Long-term unemployment in the US is worse now than iteB    has been during the past four recessions, according to a reportG    released today by the National Employment Law Project (NELP) and the $    Economic Policy Institute (EPI).      	    [snip]t         Among its key findings:  A      * Long-term unemployment has risen faster during the currenttH        recession than in any of the five previous recessions. The numberE        of workers unemployed for 26 weeks or more has risen from justXB        under 700,000 in March 2001 to 1.67 million in June 2002 anG        increase of more than 140 percent. In June 2002, there were morelF        long-term unemployed workers than at any time in the past eight
        years.e  E      * For unemployed workers, the recession has been much worse thanoE        previous recessions. At this point in the recession, long-termrH        unemployment comprises a larger percentage of the unemployed thanG        in the previous four recessions. In June, one in five unemployed ;        workers had been out of work for more than 26 weeks.i         F      * The numbers of long-term unemployed has yet to peak. During theG        last recession, long-term unemployment didnt peak for 19 months.k@        If the same pattern holds true for the current recession,E        long-term unemployment is likely to continue to rise into next E        year, well after the current extension to UI benefits expires.         nH      * Given the severity of the long-term unemployment crisis, extendedF        benefit policies are inadequate. As of June 1, over 1.4 millionA        workers were collecting up to 13 weeks of federal extendedsD        benefits. Nearly 900,000 of those workers have now run out ofF        benefits even though the labor market shows no signs of lasting        recovery.        eG    While hundreds of thousands have already exhausted federal temporaryhF    extended benefits, the report also predicts that without changes inH    federal policy thousands more will be left without benefits in comingA    months. And despite provisions for workers in states with high.?    unemployment to qualify for an additional 13 weeks, only twotE    states Washington and Oregon currently qualify for the program..."d    p    2 --Jerry Leslie   (my opinions are strictly my own)9   Note: leslie@jrlvax.houston.rr.com is invalid for email    ------------------------------  % Date: Tue, 15 Oct 2002 23:39:54 -04000- From: JF Mezei <jfmezei.spamnot@videotron.ca>. Subject: Re: Life after VMS?, Message-ID: <3DACDF81.A7B9D723@videotron.ca>   Didier Morandi wrote:r > K > I may be soon in the process of visiting a laaarge aircrafts manufacturer 
 > company   L Since your headers say you're in Toulouse, I wonder whom that large aircraftH manufacturer might be? Perhaps you could convince them to base their 380J cockpit on a Alpha cpu running VMS and then they could continue to produce/ alpha chips when HP srops support ? :-) :-) :-)i  R > 1. the "Wait&See" solution. Do nothing. As long as you have support, spare partsD > and no dysfunctioning (sp?), there is no hurry to change anything.  8 The customer must ask themselves the following question:  F Are you happy with Digital/Compaq/HP, the types of decisiosn their topM management makes ? Are they confident that HP's commitments will not change ?"( (as happened with Alpha under Compaq). ?  N If the customer feels betrayed by the vendor, then they can think of migrating right away to another vendor."  N If the customer is happy with the vendor, then they have no choice but to waitN and see. All of their enterprise products are undergoing metamorphosis and andK until that is complete and they have "completed" products on real hardware,uP then and only then can the customer really compare what the vendor has to offer.  N Right now, all of HP's enterprise product promises are vapourware because theyM will migrate to hardware that doesn't exist yet, and in the case of HP's unixfN offering, it will also undergo sighificant transformation when bits from Tru64 are integrated.z  K I'd say that if the customer intends to stay with HP, then it should have a I long chat with HP, and say all sorts of negative things, feel betrayed by-K broken commitments on Alpha, etc etc. GOAL: secure some future discounts in?H exchange for migrating to one of HP's core enterprise products when IA64M becomes a reality with some garanteed performance metric (eg: HP will providegJ as many IA64 CPUs to match what Alpha *would* have been had Alpha not beenM killed) *AND* if more CPUs are needed, get HP to compensate for the fact that L some 3rd party products will charge higher licencing  due to more CPUs being& needed to do the same job as an Alpha.  M *IF* HP's real strategy is, like Stallard has said originally, to migrate alloD VMS customers to HP-UX, then it will be to HP's advantage to provideJ singnificant enough incentives to keep those customers until HP-UX on IA64F (with True643 bits)  becomes a reality AND then convert them to HP-UX.   ------------------------------  % Date: Wed, 16 Oct 2002 00:34:11 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>l Subject: Re: Life after VMS?, Message-ID: <3DACEC36.F3E344A3@videotron.ca>   Didier Morandi wrote: Q > that DEC folks within CPQ was not that easy to reach and get involved. Now, HPQ Q > folks who have heard about VMS are really today in a shrinking handkerchief :-)eP > and there is no more VMS advisory support over here. More offline if you wish.  L This is actually an important point. In the days of Digital, the quality andL level of support from a local sales office varied greatly depending on whereK you were. Perhaps some geographies had more consistent quality across their 5 offices, but other genographies has huge differences.m  L When Compaq purchased the much larger Digital, many (if the the majority) ofL local sales offices remained "Digital staffed" since Compaq had little or no presence there.-  M The question now becomes: how long before a valid judgement can be made *by aaK customer* about the impact of the merger on the local sales office ? If oneiJ has issues with the ex Digital local sales office, then this issue becomesK important. Any chance that HP would be able change corporate culture of the  ex-Digital folks ?   ------------------------------  % Date: Wed, 16 Oct 2002 00:47:41 -0400e- From: JF Mezei <jfmezei.spamnot@videotron.ca>h Subject: Re: Life after VMS?+ Message-ID: <3DACEF5F.28ADF66@videotron.ca>    someone wrote:B > I would trust the DII-COE committment to the US Government more.  M When I sa a presentation by Rich marcello some time ago, it was revealed that.K the "commitment" is made on an individual basis to garantee support for then systems they are buying. g  N In other words, if you integrate VMS 7.2 on an aircraft carrier to control theN air conditioning system, it means that Compaq would commit to ensure that thisK system would remain operational for X years. I do not think that this means L that new releases will be made available to that customer. When you considerK how integrated many defense systems are, sometimes new releases of software-E are not desired unless really necessary since they would result in anBA extensive/expensive project to update/test/certify that software.n  G There is a part of DII/COE that has value to mortal customers: the UnixcE compatibility thing which should benefit all users of VMS eventually.e   ------------------------------  % Date: Wed, 16 Oct 2002 07:34:49 +0200-- From: Didier Morandi <Didier.Morandi@Free.fr>6 Subject: Re: Life after VMS?' Message-ID: <3DACFA78.852561F6@Free.fr>H  I I'm more and more convinced that you are a DECPQ manager, John Smith. You@  describe so well the insides :-)   John Smith wrote:2 >  ../..8N > Great idea. Be prepared to sit in your office all day because you won't haveJ > a budget for calling on prospective *new* customers, nor will you have aI > telephone to call them with, nor will you have a budget for demo units,l > etc...  M John, I learned at DEC that as long as you (try to) do the right thing, whichrL means right against you conscience, your Customers and your employer, and asI long as you *also* start bringing some $$, you get what you ask. And I am1D convinced that there are smart people at HP who can understand that.  yJ > But think of the positives -  it won't be stressful because you won't beM > competing against Sun and IBM for business. There won't be any deadlines toeL > meet for RFP's or important presentations to customers. You could probablyM > sit on the beach in Menton all day long and still meet your sales quota. HPc> > probably wouldn't even know that you were out of the office.  O Menton is the Papies and Mummies area of France, where there is 365 days a yeariO a micro-climate which is more than pleasant for their arthrite artryte artrythenL well... articulations :-) I am only 51, you see. I will probably rather move- HPEC EMEA HQ from Geneva to Sophia Antipolis.   O Wait a minute. After all, this IS a great idea. I should start to hire my team.I   D. -- y2   ------------------------------------------------2 MORANDI Consultants  http://Didier.Morandi.Free.fr0   19 chemin de la Butte, 31400 Toulouse, France.2 Tel.: +33 (0)6 7983 6418 - Fax: +33 (0)5 6154 19282 OpenVMS, APPLE, Computer Security, Migration plans2 --------------------------------------------------   ------------------------------  % Date: Tue, 15 Oct 2002 23:14:11 -0400-- From: JF Mezei <jfmezei.spamnot@videotron.ca> 0 Subject: Re: Ncurses porting and cbreak(ed) read, Message-ID: <3DACD97C.E224FC7F@videotron.ca>   Shane Smith wrote: > H > It isn't too hard in QIO, from memory all you need is a 1ch buffer andH > set it up to return when the buffer is full. You do need a little codeI > to cope with keypresses that generate multiple bytes, but that's pretty  > easy too.e  I The terminal driver is capable of handling multi-byte terminators. The IOtM status block contains and offset into the buffer where the terminator begins.m( It contains a total count of charcaters.  M The IO status buffer contains numbe of bytes read, number of terminator bytesi4 read, abnd offset of the terminator into the buffer.  N So if you enter "A" followed by PF10, (assuming PF10 generates 4 bytes), you'dL get a total of 5 bytes in the buffer, the "offset to terminator" would pointM to byte 1, you'd have a terminator size of 4, so you can deduct that you haveeL one byte of data followed by 4 bytes of terminator from the IO status block.  N The question the becomes: can the above work if you specify a buffer size of 1M for the read operation ?  (i.e. will QIO deposit a 4 byte terminator into theeM user buffer, with terminator offset of 0, if you secified a read of 1 bytes ?    ------------------------------  % Date: Tue, 15 Oct 2002 22:43:05 +0200 " From: "Hans Vlems" <hvlems@iae.nl> Subject: Re: New to VMSg5 Message-ID: <aohulg$m87c0$1@ID-143435.news.dfncis.de>e  I Hmm, my Encompass id is recognized, however I get empty email messages ina return.1. No text in the body of the msg, no attachment. Any ideas what went wrong?   Hans  ? "Phillip Sobottke" <psobottke@ameritech.net> schreef in berichtg9 news:pjEq9.4240$F53.3515143@newssvr28.news.prodigy.com...:L > It took a month from the time I joined encompass, to the time the hobbiestI > org would recognize my id.  Once my id was recognized...and I submittedt then= > license request..I had the email response within 15minutes.t- > "Joe" <syprinth@yahoo.com> wrote in message ( > news:aynq9.30460$rz6.4511@sccrnsc02...I > > I hate to come in here with all you VMS pro's, but I just got 2 alphapE > > servers and want to put VMS on them (mainly cause I don't know of  anythingE > > else that will go on them nicely).  I have applied for a hobbiestt license,I > > but have gotten no response.  Any ideas how long it takes to get one?A And.I > > 2nd, I don't want to pay the money to get the CD's.  Anyone out theree: > > willing to burn me a copy of the latest Alpha release? > > 	 > > --Joeu > >y > >a >g >D   ------------------------------  % Date: Tue, 15 Oct 2002 17:35:47 -0400 ! From: VAXVMS <bounce@notmail.com>t Subject: RE: New to VMSmK Message-ID: <BA52530E3149734A9BAABDBBFA808E4903027B84@rlghncst964.usps.gov>r  1      It's the Hobbyist Virtual Licensing program.d        WWW  I Hmm, my Encompass id is recognized, however I get empty email messages inr return. . No text in the body of the msg, no attachment. Any ideas what went wrong?   Hans  ? "Phillip Sobottke" <psobottke@ameritech.net> schreef in berichto9 news:pjEq9.4240$F53.3515143@newssvr28.news.prodigy.com... L > It took a month from the time I joined encompass, to the time the hobbiestI > org would recognize my id.  Once my id was recognized...and I submittedy thej= > license request..I had the email response within 15minutes.P- > "Joe" <syprinth@yahoo.com> wrote in message-( > news:aynq9.30460$rz6.4511@sccrnsc02...I > > I hate to come in here with all you VMS pro's, but I just got 2 alphauE > > servers and want to put VMS on them (mainly cause I don't know of- anythingE > > else that will go on them nicely).  I have applied for a hobbiest: license,I > > but have gotten no response.  Any ideas how long it takes to get one?  AndiI > > 2nd, I don't want to pay the money to get the CD's.  Anyone out theren: > > willing to burn me a copy of the latest Alpha release? > >w	 > > --Joei > >i > >m >m   ========================  William W. Webb / DSSC/RLM, USPS OpenVMS Support Services& 4924 Green Road Raleigh, NC 27616-2800: 919.874.3043 <FirstInitialDotLastNameAtEmailDotUSPSDotGov>   ------------------------------  # Date: Wed, 16 Oct 2002 01:49:23 GMTl- From: "John E. Malmberg" <wb8tyw@qsl.network>o Subject: Re: Old DECC for VAXn= Message-ID: <DA3r9.32270$m92.6210814@news1.news.adelphia.net>0   Simon Brown wrote:   > Hi John,   >iD   > The customer has an older version of AACRT??? and will not allowB   > AACRT060 to be installed, so I have to get an older version ofA   > the compiler, which a kind person is making available for me.c  C I think that there has been some confusion.  AACRT060 is the oldestr, version of AACRT*.  There are no older ones.   There are ECO kits.   ? The AACRTL060 kit may be detecting the ECO kit and aborting the ! installation of it's older image.n  F If this is the case, then the required DEC C RTL is already present on. the customer machine, and no action is needed.  C What is the exact error messages when installing the AACRTL060 kit?a  / What does dir sys$help:*CRTL*.RELEASE* produce?c  D Now for installing C programs from newer versions of the compiler onH older operating systems, there are a few issues that the AACRTL will not solve.  B The AACRTL does not supply all of the newer routines.  Some of the> routines like the UTC routines will not be present because the0 underlying support is not then in OpenVMS 5.5-2.  E Some routines can be supplied by linking with the "Backport" run timeh& library, see the file DECC$CRTL.README  E None of these problems will be fixed by installing an older compiler.t   -John  wb8tyw@qsl.network Personal Opinion Onlym   ------------------------------  % Date: Tue, 15 Oct 2002 22:19:52 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>g Subject: Re: Old DECC for VAXC, Message-ID: <3DACCCC4.B8BB2593@videotron.ca>   Simon Brown wrote:N > The customer has an older version of AACRT??? and will not allow AACRT060 toJ > be installed, so I have to get an older version of the compiler, which a) > kind person is making available for me.s  N Out of curiosity, couldn't you simply provide the object modules and link them5 on their older system with whatever C-RTL they have ?m  G I realise that the AACRTL patches will fix problems, but do they change  functionality ?i   ------------------------------  % Date: Tue, 15 Oct 2002 20:02:32 +0200t6 From: Arne =?ISO-8859-1?Q?Vajh=F8j?= <arne@vajhoej.dk>! Subject: Re: OpenVMS is strategico' Message-ID: <3DAC5838.10307@vajhoej.dk>i  ( Andrew Harrison SUNUK Consultancy wrote:  8 > In case you hadn't worked it out a STRATEGIC MARKETING6 > program that is designed just to keep the readers of4 > COV happy doesn't qualify as being strategic or an0 > effective use of marketing resources. Hence my
 > comment.     And ?n   Noone has questioned that.   But that was not the topic.    It is completely irrelevant.  4 Just a good excuse for a sales guy to post some FUD.  8 > If you have any other comprehension problems I suggest> > you refer them to a co-worker because as a public discussion > its pretty tedious.i    < Feel free to leave comp.os.vms/INFO-VAX any time, if you get too tired of reading my posts !m  0 >> Listen carefully: THE TOPIC IS NOT MARKETING,/ >> THE TOPIC IS NOT ABOUT WHY HP IS DOING THIS,p4 >> THE TOPIC IS WHETHER IT MAKES A DIFFERENCE OR US. > 2 > So explain why HP should be interested something > that flatters COV readers ???t     I do not think they are.   Noone has ever said so either.  4 > You describe my posts as stupid but in the process1 > illustrate in all to graphic detail why OpenVMSa1 > is likely to die (assuming that the majority ofg$ > OpenVMS customers agree with you).  7 I think I can understand why SUN's stock-price is where:
 it is now.  8 If SUN's sales-people think that the fact that VMS users8 will enjoy seeing a VMS brochure send out to IBM and SUN9 customers will kill VMS, then that company has a problem.e A big problem.   Arne   ------------------------------    Date: 15 Oct 2002 15:01:55 -0600- From: koehler@encompasserve.org (Bob Koehler)y! Subject: Re: OpenVMS is strategicT3 Message-ID: <Tx+9Jp1ucpLT@eisner.encompasserve.org>V  X In article <3DABFCA2.6030104@nospamn.sun.com>, Andrew Harrison SUNUK Consultancy writes: >  > ; > But not an answer that had any relevance to the question.: >    Lets do it again, shall we:P   In response to Arne Varhj:o  * > Andrew Harrison SUNUK Consultancy wrote: >e7 >So you make no distinction whatsover between effectivei. >marketing and marketing that is ineffectual ! >h
 >Unsuprising.e >e0 >This is a horribly tactical marketing programme3 >it is all about getting customers who have planneds; >OpenVMS purchases to bring them forward into this quarter.a >t< >Given its short validity and the target base it is unlikely: >to have any impact on Sun/IBM shops, unless they also buy7 >OpenVMS, have a purchase in the pipeline and happen to68 >have some decommisioned IBM/Sun boxes that they can use8 >to make HP eat some more margin. Since all of this is a; >bit of a long shot you can see that the difference betweeno5 >effective marketing and ineffective marketing is all  >important.d >t: >If you can think of one single valid reason why a Sun/IBM8 >account that does not have any OpenVMS systems would be@ >interested in this marketing programme then please advance them4 >otherwise my point which was short and to the point >stands. >  >Regards >Andrew Harrison  I    Now you asked for "one single valid reason why a Sun/IBM account that cA    does not have any OpenVMS systems would be interested in this .    marketing programme".  I    You did not ask for a reason _present within this marketing program's iF    materials_, as you have since implied.  You didn't ask it that way,;    so I'm under no obligation to reduce my answers to such.h  
    I said:   >   How about 3 valid reasons: >  >   1) securityf >   2) reliability >   3) ease of use  F    And those are 3 perfectly valid reasons for anyone to be interestedD    in any marketing campaign for said product.  Wanting to know whatF    more about that product is certainly appropriate and however littleC    information the marketing in question may provide the Sun or IBMn6    account won't know until they persue thier interst.   ------------------------------  % Date: Tue, 15 Oct 2002 21:07:58 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>t4 Subject: Re: Originator of detached process from DCL' Message-ID: <3DACC9FE.7BEA1ED2@fsi.net>-   Joshua Cope wrote: > D > On Alpha V7.3-1, there's a new GETJPI code for the parent process: > - >   $ write sys$output f$getjpi("","CREATOR")  >      20602D445   Works on V7.3, also.   Interesting bits:e  E RUN/DETACHED - "CREATOR" points to the process where the RUN/DETACHEDl command was executed.d  % SUBMIT - "CREATOR" points to the JBC.h  2 Login at the console - "CREATOR" points to the JBC  = TELNET to a system running Multinet - "CREATOR" points to the  MULTINET_SERVER process.  D I don't have access to a V7.3+ system running either TCPware or UCX.   -- i David J. Dachteraa dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/<   ------------------------------    Date: 15 Oct 2002 15:17:39 -0600- From: koehler@encompasserve.org (Bob Koehler)h3 Subject: Re: OT Humour - If Microsoft made toastersd3 Message-ID: <cVNefpQlEMMZ@eisner.encompasserve.org>e   In article <rdeininger-1410022238390001@1cust35.tnt2.nashua.nh.da.uu.net>, rdeininger@mindspring.com (Robert Deininger) writes:e > I > Why did you hard-code the toaster name?  Clearly VMS would support morem@ > than one toaster, so the "TOA0:" string should be generalized. >   >   Such generalizations are frequently ignored in sys$examples:   > Also, you forgot >    include '($toastdef)'  F   I left all the others out, too.  These were extracts.  I didn't want$   to bore you with the obvious.  8-)   ------------------------------    Date: 15 Oct 2002 15:18:33 -0600- From: koehler@encompasserve.org (Bob Koehler)F3 Subject: RE: OT Humour - If Microsoft made toastersg3 Message-ID: <BSzSAJjt+EiA@eisner.encompasserve.org>w  o In article <BA52530E3149734A9BAABDBBFA808E4903027B80@rlghncst964.usps.gov>, VAXVMS <bounce@notmail.com> writes:k > 5 > But TOA0: is just a template (like LTA0:) isn't it?e > K > Terminal TOA0:, device type unknown, is unplugged, heated, slice-orientedu: >     device, darkness control, device is a template only. > @ >     Error count                    0    Operations completed 0< >     Owner process                 ""    Owner UIC [SYSTEM]D >     Owner process ID        00000000    Dev Prot S:RWPL,O:RWPL,G,W@ >     Reference count                0    Default buffer size 80      Oh, yes!/   ------------------------------  % Date: Tue, 15 Oct 2002 22:48:46 +0200t" From: "Hans Vlems" <hvlems@iae.nl>3 Subject: Re: OT Humour - If Microsoft made toastersM5 Message-ID: <aohv07$mp08s$1@ID-143435.news.dfncis.de>-  < "Bob Koehler" <koehler@encompasserve.org> schreef in bericht- news:BSzSAJjt+EiA@eisner.encompasserve.org...( > In articleG <BA52530E3149734A9BAABDBBFA808E4903027B80@rlghncst964.usps.gov>, VAXVMS  <bounce@notmail.com> writes: > >n7 > > But TOA0: is just a template (like LTA0:) isn't it?  > > > > > Terminal TOA0:, device type unknown, is unplugged, heated, slice-oriented< > >     device, darkness control, device is a template only. > >TB > >     Error count                    0    Operations completed 0> > >     Owner process                 ""    Owner UIC [SYSTEM]F > >     Owner process ID        00000000    Dev Prot S:RWPL,O:RWPL,G,WB > >     Reference count                0    Default buffer size 80 >v
 >    Oh, yes!l > I Future release suggestion: could I have cluster support for TOA0: since IrH work for a large catering company and we want to have our toast done the same way....   ------------------------------  % Date: Tue, 15 Oct 2002 23:59:04 -0400s- From: JF Mezei <jfmezei.spamnot@videotron.ca> 3 Subject: Re: OT Humour - If Microsoft made toasters , Message-ID: <3DACE3FE.C1334763@videotron.ca>  
 VAXVMS wrote:o > 5 > But TOA0: is just a template (like LTA0:) isn't it?n > K > Terminal TOA0:, device type unknown, is unplugged, heated, slice-orientedt: >     device, darkness control, device is a template only.  N But then, you'd do an ASSIGN to TOA0:, then you'd have to use $QIO IO$M_ACCESSE ,to connect to the right toaster. At that point, you may need to do anN IO$M_SENSEMODE to find out more about that particular toaster (how many slicesG can it do at once ?, how fast does it work ? Does it support additionalt. function modifieres such as IO$M_EMPTYCRUMBS ?  N You'd also need a special function code to cause an AST to be delivered shouldF the user press the "EJECT BUTTON" on the unit, so that your program isM notified that the toasting was interrupted. (or perhaps just use the standard  CTRL-Y/C interrupt handling ?)   ------------------------------  # Date: Wed, 16 Oct 2002 01:37:05 GMTl- From: "John E. Malmberg" <wb8tyw@qsl.network>o2 Subject: Re: OVMS Concurrent user (update license)= Message-ID: <5p3r9.32244$m92.6209382@news1.news.adelphia.net>-   KCLim wrote:  > Hi,  >=  > Can someone please clarify for me what is update license..t  C Generally an update license is an enhancement or replacement for ant existing license..  D  > I wanna migrate from VAX to Alpha and it is my understanding thatC  > the OVMS user license can be upgraded to the Alpha OVMS. Is thisrD  > update license? If not, what is. I can't seem to find anything on  > the OVMS SPD.  C The available upgrades for existing licenses are a function of both ? standard terms and conditions, and various marketing functions.n  B You will want to contact an HP sales representative to discuss the- options that are open to you for the upgrade.0  H Many of the hardware purchase programs have generous trade in or upgrade terms for existing licenses.     -John. wb8tyw@qsl.network Personal Opinion Onlyt   ------------------------------  % Date: Tue, 15 Oct 2002 21:21:44 -0500o1 From: "David J. Dachtera" <djesys.nospam@fsi.net>  Subject: Re: PeopleSoft on VMS' Message-ID: <3DACCD38.EB3D0DFF@fsi.net>t   Tony Scandora wrote: > > > "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message# > news:3DAB8B04.EE7A3E58@fsi.net...a > > David Rabahy wrote:i > > >iM > > > If we can build a compelling business case then perhaps PeopleSoft wills > be > > > persuaded. [snip]d > >aE > > *SIGH* What part of "Customers want it" do they find difficult totI > > understand? At this point, "compelling business case" is as simple asoD > > "HP-UX is not OpenVMS, and HP owns both". If HP goes in with theH > > attitude, "Alright, what will it take to make this happen?" I'm sureJ > > we'll see it happen - but frankly, and this is admittedly "sentiment",I > > neither is likely: HP is not likely to do that, so it's not likely ton > > happen.  > N > The real question is not qualitative customers want it, but quantitative how > many customers.  Ditto SAP.   D ...and as I posted further on, doing that without actually trying to> sell a propsect on a non-existent product is rather difficult.   -- e David J. Dachtera  dba DJE Systemsu http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/E   ------------------------------  % Date: Tue, 15 Oct 2002 21:24:02 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>r Subject: Re: PeopleSoft on VMS' Message-ID: <3DACCDC2.B97863C3@fsi.net>M   Shane Smith wrote: > I > I wonder if "it doesn't matter if anyone wants it, you have to do it toSJ > give people some confidence in your support for the platform" would fly? >  > Yeah, right, I'm dreaming.  C A wise man once told me that even if only one customer asks for it,:G *THAT* is a market ... because he'll tell his colleagues, who will tell " their colleagues, who will tell...   -- i David J. DachteraM dba DJE Systemsr http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/n   ------------------------------  % Date: Tue, 15 Oct 2002 17:42:05 -0400 ! From: VAXVMS <bounce@notmail.com>n, Subject: Question about f$getsyi and ItaniumK Message-ID: <BA52530E3149734A9BAABDBBFA808E4903027B85@rlghncst964.usps.gov>-  % I'm writing some forward-looking DCL.c  3 f$getsyi("ARCH_TYPE") returns 1 if VAX, 2 if Alpha.m  5 Will the return value be 3 for Itanium-based systems?x   ========================  William W. Webb / DSSC/RLM, USPS OpenVMS Support Services& 4924 Green Road Raleigh, NC 27616-2800: 919.874.3043 <FirstInitialDotLastNameAtEmailDotUSPSDotGov>   ------------------------------  % Date: Tue, 15 Oct 2002 21:55:35 -0400t2 From: rdeininger@mindspring.com (Robert Deininger)0 Subject: Re: Question about f$getsyi and ItaniumK Message-ID: <rdeininger-1510022155350001@1cust222.tnt3.nashua.nh.da.uu.net>e  
 In articleG <BA52530E3149734A9BAABDBBFA808E4903027B85@rlghncst964.usps.gov>, VAXVMSc <bounce@notmail.com> wrote:o  & >I'm writing some forward-looking DCL. >n4 >f$getsyi("ARCH_TYPE") returns 1 if VAX, 2 if Alpha. >r6 >Will the return value be 3 for Itanium-based systems?  2 Very likely, but an official answer would be nice.  J If you have access to software service, you could place a call and ask for2 an official answer.  It may be too early, however.   ------------------------------  % Date: Wed, 16 Oct 2002 00:54:49 -0400e- From: JF Mezei <jfmezei.spamnot@videotron.ca> 0 Subject: Re: Question about f$getsyi and Itanium, Message-ID: <3DACF10B.CE81D23C@videotron.ca>   Robert Deininger wrote:u8 > >Will the return value be 3 for Itanium-based systems? > 4 > Very likely, but an official answer would be nice.  L Wouldn't it perhaps be logical to have it as "4" since it would then presentJ the opportunity for bits (if bit 0 is set, then vax, if bit 1 is set, then& alpha, if bit 2 is set than IA64 etc).  N However, at this point of time, wouldn't such a decision have been made a long" time ago and already implemented ?   ------------------------------  % Date: Tue, 15 Oct 2002 15:15:29 -0400-; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com> ' Subject: Re: singular and plural of VAXf$ Message-ID: <3dac6982$1@news.si.com>  ( >That is true, but for different things.  > No, not for different things, but as alternates of each other. -- tA Brian Tillman                   Internet: tillman_brian at si.comsA Smiths Aerospace                          tillman at swdev.si.come= 3290 Patterson Ave. SE, MS      Addresses modified to preventt< Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  % Date: Tue, 15 Oct 2002 22:08:12 -0400m- From: JF Mezei <jfmezei.spamnot@videotron.ca>-+ Subject: Re: store .MOV file on a VMS disk?r, Message-ID: <3DACCA09.6275113C@videotron.ca>   Elliott Roper wrote:G > it. I get slightly better on the way back, about 900KB/sec. The rounduI > trip for a 55 minute movie would take 10 hours and eat 13GB of VMS diske > space.  M Depends on what compression you use for the quicktime movies. (and of course,b> the size of the images, sound quality and number of channels).    V On the VMS system, you need to ensure that TYPE IMAGE is used (eg: raw data transfer).  K On the MAC, an FTP client(such as Fetch)  allows you to specify whether you N want macbinary or raw binary. Macbinary will sedn , as one file, both the dataL and resource forks, so that on the return trip, the file is recreted exactlyM the same. But if you wish to publish the .MOV to the net, then you should useaK the raw binary transfer to strip off the resource fork. (in both cases, VMSyC doesn't see a difference, it just receives and stores binary data).t   ------------------------------    Date: 15 Oct 2002 21:36:19 -0700: From: craig.berry@SignalTreeSolutions.com (Craig A. Berry)+ Subject: stuck process after v7.3-1 upgradee= Message-ID: <7f15589f.0210152036.3ba33555@posting.google.com>R  F I'm seeing a readily reproducible issue that showed up after upgradingF from v7.3 to 7.3-1 where a process consumes a lot of CPU, about 75% inD kernel mode, and claims to be doing 50,000-70,000 DIOs per second on5 my DPW 500au. Yes, you read that right; MONITOR, SHOWiE PROCESS/CONTINUOUS, and Ctrl-T all think this process is doing I/O atnF a preposterous rate, though SDA shows that none of the channels it hasC open is busy and the disk is quiet. If I'm interpreting them right,hA the PC values from SHOW PROCESS/CONTINUOUS suggest it is bouncinge2 around in P2 space (800FF940, 800B62D0, 801021F8).  ? The problem occurs with or without XFC enabled, with or withoutoC operating from an ODS-5 disk (system or otherwise), with or withoutr@ hard links enabled, with or without VMS731_RMS V1.0, and with orD without VMS731_SYS V1.0. SDA shows no evidence of quotas being maxed out.  @ In addition to my DPW 500au, others report identical behavior on? AlphaServer 400 and AlphaServer 800; I don't think any of thesei/ systems is a candidate for the newest firmware.g  C One person reports not being able to reproduce the problem; I don'ttE know what system he had but he mentioned fibre channel disks, whereasI/ the systems that exhibit the problem have SCSI.g  D Enabling SET WATCH FILE makes the symptoms disappear -- possibly the/ bug, if that's what it is, is timing sensitive?   E Peeling back the stack of the stuck process by issuing SHOW CALL/NEXTr2 within the SDA (while logging the output) gives me  ! $ search sda.log "return address"n2  Return address on stack = FFFFFFFF.800B6468       EXE$CMODKRNL_C+001982  Return address on stack = 00000000.0004D090       ESIS$DEFINE_PPI+00CD0hA  Return address on stack = FFFFFFFF.8014D0D0       EXE$AST_RETURNr,  Return address on stack = 00000000.7AE57230E  Return address on stack = 00000000.00043FC0       CTF$SRV$INIT+00A80e<  Return address on stack = 00000000.00041714       MMK+417142  Return address on stack = 00000000.00040068       UCB$M_SUPMVMSG+00068,  Return address on stack = 00000000.7AF7FDD8,  Return address on stack = 00000000.7AF7FC18  @ To reproduce the problem, download the MMS replacement MMK from    <http://www.madgoat.com/mmk>  > Create a directory, unzip, link, and define "MMK" as a foreign command.  E Sorry not to have a smaller reproducer, but you'll then need to get ar Perl distribution from  ' <http://www.cpan.org/src/stable.tar.gz>   @ Unpack the archive (requires gunzip and vmstar).  Build like so:   $ set default [.perl-5_8_0]d& $ @configure -"des"  ! quotes required    . . .   $ mmkv  C The core build goes fine but once it starts building the extensionsl? the subprocess running MMK gets stuck. Perl is just an innocent A bystander here; it happens to have rather complicated descriptionnB files that manage to trigger the problem. I suspect MMK is also anC innocent bystander but am less sure. But even if MMK does somethingeE wrong (and it worked just fine under v7.3), there is clearly a deepercB problem if the OS thinks a process is doing > 50K DIOs per second.  F This issue has been previously discussed on the vmsperl mailing list. C Go to <http://www.xray.mpe.mpg.de/mailing-lists/vmsperl/> and entern4 "mmk; 7.3-1" in the search box (without the quotes).  @ My 7.3-1 system is a hobbyist one so this is not something I can@ report through channels. If anyone can confirm these symptoms or suggest a fix I'd be grateful.   ------------------------------  # Date: Tue, 15 Oct 2002 21:41:53 GMTa' From: Rick Dyson <rick-dyson@uiowa.edu>- Subject: Tape Copy?3( Message-ID: <3DAC8BA1.653FA3A@uiowa.edu>  D I need to duplicate a DLT I have with many backup save-sets on it toC another DLT.  It would be nice if the creation date of the new copywE save-sets were preserved from the original, but that is not critical.d  . The save-sets were made with /BlockSize=65535.  - I have tired to use COPY, but get this error:o  ( $ Mount /Block = 65535 MISC$MKD600: dpm1/ %MOUNT-I-MOUNTED, DPM1 mounted on _MISC$MKD600:o0 $ Mount /Block = 65535 /NoWrite fvs2$MKc600: dpm. %MOUNT-I-MOUNTED, DPM mounted on _FVS2$MKC600:+ $ Copy /Log FVS2$MKC600:*.* MISC$MKD600:*.*sH %COPY-E-OPENIN, error opening FVS2$MKC600:[]DPM0101-011005.BC;1 as input@ -RMS-F-IRC, illegal record encountered; VBN or record number = 0  C I have also tried to use BACKUP to copy the files from tape to tapel without any success.  ! What am I missing or doing wrong?h   Rick   ------------------------------  % Date: Tue, 15 Oct 2002 21:53:35 -0400e2 From: rdeininger@mindspring.com (Robert Deininger) Subject: Re: Tape Copy? K Message-ID: <rdeininger-1510022153350001@1cust222.tnt3.nashua.nh.da.uu.net>n  J In article <3DAC8BA1.653FA3A@uiowa.edu>, Rick Dyson <rick-dyson@uiowa.edu> wrote:  E >I need to duplicate a DLT I have with many backup save-sets on it to D >another DLT.  It would be nice if the creation date of the new copyF >save-sets were preserved from the original, but that is not critical. >o/ >The save-sets were made with /BlockSize=65535.i >a. >I have tired to use COPY, but get this error: >t) >$ Mount /Block = 65535 MISC$MKD600: dpm1s0 >%MOUNT-I-MOUNTED, DPM1 mounted on _MISC$MKD600:1 >$ Mount /Block = 65535 /NoWrite fvs2$MKc600: dpme/ >%MOUNT-I-MOUNTED, DPM mounted on _FVS2$MKC600:e, >$ Copy /Log FVS2$MKC600:*.* MISC$MKD600:*.*I >%COPY-E-OPENIN, error opening FVS2$MKC600:[]DPM0101-011005.BC;1 as inputoA >-RMS-F-IRC, illegal record encountered; VBN or record number = 0r >dD >I have also tried to use BACKUP to copy the files from tape to tape >without any success.  >m" >What am I missing or doing wrong?  G Backup wants tapes mounted /FOREIGN.  You can usually let BACKUP do thei	 mounting.m  " Tape-to-tape copy is a bit tricky.  J The best way to copy backup tapes is via the Save Set Manager product.  IfD you have the campuswide licenses and an SPL media kit, load save setG manager and use it.  If you don't have a license, consider requesting aeC free 30-day trial PAK.  SSM can change the blocking, CRC, and othereI properties of save sets, and can back up to 5 output copies in parallel.  F If you are serious about backups and value your time, this is likely a product worth buying.e  I I recommend limiting tape save sets to /BLOCK=32256, which is the largest H block size allowed for disk save sets.  These save sets can be COPY'd toH disk (not as reliable as Save Set Manager, but often still useful) where9 they are more convenient to sort through in an emergency.u   ------------------------------  % Date: Tue, 15 Oct 2002 23:17:52 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> 9 Subject: Re: TCPIP, sharing sockets via $QIO (fixed yet?)e, Message-ID: <3DACDA58.4E6F13DB@videotron.ca>   Mark Daniel wrote: > F > Back in March this year I initiated a thread about sharing listening' > sockets under Compaq TCP/IP Services.l  K > I have been watching the TCP/IP releases and ECOs since but have not beeni> > able to find any reference to this issue.  Have I missed it?  M Summer is arriving in Australia, so the TCPIP engineers on the gold coast areuL spending more time on the beach waxing their surfboards  :-) :-) :-) :-) :-) :-) :-) :-) :-)    ------------------------------  # Date: Wed, 16 Oct 2002 01:07:12 GMTa- From: "John E. Malmberg" <wb8tyw@qsl.network>u2 Subject: Re: TCPIP,VMS 7.2.1 Alpha, license issue?= Message-ID: <4Z2r9.32192$m92.6204681@news1.news.adelphia.net>F   Brian wrote:  > Thanks up front.hH  > We've got a DEC 2100 Alpha with VMS 7.2.1 that we are trying to bring;  > up.  Once up we need to FTP some files to a Windows box. <  > We are getting an error when trying to FTP from that box..  > "...OPERATION REQUIRES SOFTWARE LICENSE..."  >!  > When we check licensing, I seea  > NET-APP-SUP-200  # How are you checking the licensing?n  < Some times temporary license keys are installed on a system.F If this system was a member of a cluster, then the license may need to# be assigned to the specific system.m  F The NET-APP-SUP-200 license key provides full TCP/IP functionality and quite a few other products.a    > UCX  D The UCX license key is redundant to the NET-APP-SUP-200 license key.? Provided that they are not expired and are installed correctly.r  
  > Any ideas?i  G The documentation for OpenVMS, would be of help to you, and also pleaserG consult the OpenVMS FAQ section on trouble shooting license key issues.a  A Both are available from links from http://www.openvms.compaq.com/u   -JohnM wb8tyw@qsl.network Personal Opinion Only-   ------------------------------  # Date: Wed, 16 Oct 2002 01:14:03 GMTi- From: "John E. Malmberg" <wb8tyw@qsl.network>M2 Subject: Re: TCPIP,VMS 7.2.1 Alpha, license issue?= Message-ID: <v33r9.32205$m92.6205090@news1.news.adelphia.net>h   Brian wrote:   > Thanks up front.I   > We've got a DEC 2100 Alpha with VMS 7.2.1 that we are trying to bringo<   > up.  Once up we need to FTP some files to a Windows box.=   > We are getting an error when trying to FTP from that box.h/   > "...OPERATION REQUIRES SOFTWARE LICENSE..."    >e"   > When we check licensing, I see   > NET-APP-SUP-200n  # How are you checking the licensing?h  < Some times temporary license keys are installed on a system.F If this system was a member of a cluster, then the license may need to# be assigned to the specific system.t  F The NET-APP-SUP-200 license key provides full TCP/IP functionality and quite a few other products.      > UCXe  D The UCX license key is redundant to the NET-APP-SUP-200 license key.? Provided that they are not expired and are installed correctly.a     > Any ideas?  G The documentation for OpenVMS, would be of help to you, and also pleaseeG consult the OpenVMS FAQ section on trouble shooting license key issues.   A Both are available from links from http://www.openvms.compaq.com/e   -Johni wb8tyw@qsl.network Personal Opinion Onlys   ------------------------------  % Date: Tue, 15 Oct 2002 22:19:59 +0200i" From: "Hans Vlems" <hvlems@iae.nl>0 Subject: Re: Telnet problem after 7.3-1 Upgrade.5 Message-ID: <aohtaa$mnuki$1@ID-143435.news.dfncis.de>I  K > How could I best check this, & how could I disable it. I have a couple of-J > third party companies use our system, so wouldn't have their PC's in anyJ > sort of DNS. I just want what I had before, telnet connecting in < 1 sec6 > without caring where the connection was coming from. > >SL Use nslookup to figure out the problem (be sure to read the book on BIND andI DNS by Cricket Liu, DNS is getting important). My systems are in a domain K called vlems.thuis and the systems are called helium, argon etc. An exampleaC of what nslookup returns on my VAX/VMS V7.2 and TCPIP V5.0 systems:h  H The following session was recorded on node ARGON, my DNS runs on HELIUM:   $ mc tcpip$nslookupeK Default Server:  helium.vlems.thuis                                     [1]p Address:  192.168.44.1   > vlems.thuis. [2]  Server:  helium.vlems.thuiso Address:  192.168.44.1   Name:    vlems.thuis   > ls vlems.thuis.r [3]e [helium.vlems.thuis] $ORIGIN VLEMS.THUIS.5 chloor                  12H IN A        192.168.44.10 4 argon                   12H IN A        192.168.44.64 xenon                   12H IN A        192.168.44.24 vanadium                12H IN A        192.168.44.74 fluor                   12H IN A        192.168.44.84 helium                  12H IN A        192.168.44.15 jodium                  12H IN A        192.168.44.11o4 radon                   12H IN A        192.168.44.34 krypton                 12H IN A        192.168.44.54 broom                   12H IN A        192.168.44.94 neon                    12H IN A        192.168.44.4 > set q=soaz [4]1 > vlems.thuis. [5]i Server:  helium.vlems.thuisa Address:  192.168.44.1   vlems.thuise#         origin = helium.vlems.thuisu-         mail addr = SYSTEM.helium.vlems.thuis-         serial = 13a         refresh = 3600 (1H)          retry   = 300 (5M)         expire  = 172800 (2D)w!         minimum ttl = 43200 (12H)$/ vlems.thuis     nameserver = helium.vlems.thuis 7 helium.vlems.thuis      internet address = 192.168.44.1  >u  L [1] this message is returned by nslookup; it tells what DNS server it found.L [2] just enter the domain name, terminate with a PERIOD (it's like COBOL :-)=      and basically the same message as under [1] is returned.tK [3] shows what nodes are registered in the domain, the command ls does justr that; remember the      period. L [4] set query=soa (start of authority); changes the type of query of the DNSI [5] typing the domain name+period lists all kind of information about the  DNS server settings.   Hope this helps.   Hans   ------------------------------    Date: 15 Oct 2002 15:04:46 -0600- From: koehler@encompasserve.org (Bob Koehler) 7 Subject: Re: Unix file descriptors and VMS I/O channelsn3 Message-ID: <en3hljJ7fYWB@eisner.encompasserve.org>s  c In article <3DAC05B1.44ABD37A@mindspring.com>, Atlant Schmidt <atlantnospam@mindspring.com> writes:w > Larry Kilgallen wrote: > % > And if you omit SYS$, don't forget:  >  >   RSX-11M   B    It was certainly SYS$QIO on my copy of RSX-11M.  Of course, youD    could use the $QIO macro in assembly, but you can do that on VMS,    too.    ------------------------------  % Date: Tue, 15 Oct 2002 23:33:51 +020006 From: Arne =?ISO-8859-1?Q?Vajh=F8j?= <arne@vajhoej.dk>7 Subject: Re: Unix file descriptors and VMS I/O channels ) Message-ID: <3DAC89BF.3030000@vajhoej.dk>z   Z wrote:  5 > Johan Nilsson <johan.nilsson@esrange.ssc.se> wrote:bF > : programming question: Is it possible to get the underlying VMS I/OI > : channel no from an UNIX style file descriptor? I'd like to be able to4E > : use open() to get valid file descriptor, and then use QIO for I/Ot > : operations on these files. > G > If you're going to go through all the trouble of using $QIO to write .< > to the file, why not also use $ASSIGN to get the channmel?    0 Becaue $ASSIGN only opens a channel to the disk.  $ You need the FID to access the file.   That is why it is common to:
    $OPEN filee    save FIDo    $CLOSE file    $ASSIGN channel to disk    $QIOW access file via FID    $QIOW write to file    $QIOW deaccess file    $DASSGN channel   Arne   ------------------------------  % Date: Tue, 15 Oct 2002 21:12:02 -0000r! From: Z  <zarlenga@conan.ids.net>e7 Subject: Re: Unix file descriptors and VMS I/O channelsn/ Message-ID: <uqp152palp0p9b@corp.supernews.com>n  3 Johan Nilsson <johan.nilsson@esrange.ssc.se> wrote: D : programming question: Is it possible to get the underlying VMS I/OG : channel no from an UNIX style file descriptor? I'd like to be able to C : use open() to get valid file descriptor, and then use QIO for I/O> : operations on these files.  E If you're going to go through all the trouble of using $QIO to write b: to the file, why not also use $ASSIGN to get the channmel?   ------------------------------  # Date: Wed, 16 Oct 2002 04:08:04 GMTt- From: "John E. Malmberg" <wb8tyw@qsl.network> 7 Subject: Re: Unix file descriptors and VMS I/O channels0= Message-ID: <EC5r9.32548$m92.6235002@news1.news.adelphia.net>    Johan Nilsson wrote: > Hi,s > D > programming question: Is it possible to get the underlying VMS I/OG > channel no from an UNIX style file descriptor? I'd like to be able to C > use open() to get valid file descriptor, and then use QIO for I/Ot > operations on these files.  D Intermixing SYS$QIO and RMS access to files opened by the C RTL can B cause the file to get corrupted because the C RTL buffers the I/O.  C If you really want to do this, and you are sure you understand the 0E consequences, then you can use the rms acc callback extension on the nK open() call to get at the RMS RAB and FAB structures used to open the file.y    From there it is a SMOP.   F However it might be safer to write your own wrapper module for open() E and close() that do the VMS specific code and leave the C RTL out of e+ things if you want to do the I/O your self.o   -John  wb8tyw@qsl.network Personal Opinion Onlyl   ------------------------------  % Date: Tue, 15 Oct 2002 14:07:53 -0400T& From: David M Smith <dsmit115@csc.com>& Subject: Re: VMS 5.3 disk installation' Message-ID: <aohlhq$n4v$1@lore.csc.com>l  F On Tue, 15 Oct 2002 11:10:29 -0400, John <cmiinkNOSPAM@msn.com> wrote:   > N > I have a VAX 3200 with VMS 5.3 . My knowledge on Unix, VAX 3200's and VMS isM > very lacking. I am trying to change the store drive from a Seagate ST12550N  > (2 GB) to a ST15150N (4 GB).K > When formatting I get options Node, logical unit and INTERLEAVE (0-7).  IgM > don't know what Interleave is, so I chose zero (0), and the formatting went L > OK. When bringing up the system and initializing the new drive I get errorN > (INIT-F-Cluster, unsuitable cluster factor).  Can anyone help configure thisL > drive into the system?  I don't  know if this drive is suitable, or how toA > change the cluster size.  I would appreciate any help or input.: > Thanks.. John Pittman=  E I don't know of any reason this won't work, and Nic's suggestions mayeH allow you go get it initialized. I recall having to specify "reasonable"C cluster sizes back in that time frame as disks began to get larger.R  H Do be aware of a hard-and-fast limit on the maximum size of a disk whichE can be connected to V5.5-2 (and earlier) systems, which, if exceeded,rF can result in data loss and/or disk corruption with no obvious errors.A This limit was raised in VMS V6.0, and is documented (only, to my H knowledge) in section 2.2.21 of the VMS V6.0 Release Notes (not a widely read document around here):I  /   2.2.21 Volume Size-Maximum Increased to 2**31s   V6.0  9   In previous versions of OpenVMS, the maximum size of anc9   individual volume supported by the file system was lim-i<   ited to 2**24 blocks (approximately 16,000,000 blocks or 88   GigaBytes). In Version 6.0, the maximum supported vol-<   ume size has been increased to 2**31 blocks (approximately>   2,000,000,000 blocks or 1 TeraByte.) While there are no sin-A   gle disks of this size available, it is possible to construct a C   logical volume of this size using striping or similar techniques.   =   Note that in a volume set, the size limit applies individu-d=   ally to each member of a volume set. Thus, in theory, it isiA   now possible to construct a volume set with a total size of 255:   TeraBytes.E ---------------------------------------------------------------------sE David M. Smith 302.391.8533                   dsmit115 at csc dot comtE Computer Sciences Corporation (Opinions are those of the writer only)-E ---------------------------------------------------------------------n   ------------------------------  # Date: Wed, 16 Oct 2002 03:08:21 GMTg. From: "Tom Simpson" <simpsont@attbi.com.fubar>9 Subject: VMS 7.3-1 breaks clustering using memory channelr- Message-ID: <FK4r9.59722$rz6.10542@sccrnsc02>o  D We recently upgraded from VMS 7.2-1 to 7.3-1.  After the upgrade, we
 discoveredJ that clustering was no longer possible, since VMS 7.3-1 does not work with
 the memoryJ channel interface on ES4x and DS20 systems (says HP support).  This is not an intermittenteL or hard to reproduce problem.  The memory channel fails to initialize during the boot andH unless you have an alternate cluster path enabled, you are screwed.  Our alternate path is G a redundant memory channel interface and hub, which is no help for thisn
 situation.  I This makes me REAL nervous about the quality control or VMS these days...h  J I'm surprised that no one else on this NG has run into this problem before now.  C HP says VMS 7.3 works ok, but 7.3-1 does not (which I can confirm).t   Regards, Tomo   ------------------------------  % Date: Tue, 15 Oct 2002 15:27:07 -0400e' From: "Main, Kerry" <Kerry.Main@hp.com> & Subject: RE: VMS 7.3-1 File CorruptionT Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF4023D9771@kaoexc01.americas.cpqcorp.net>   Craig,   Just a thought -  F About 18 months ago, there was a flavour of the KGPSA FC HBA adapter =J that had a firmware problem that caused all sorts of issues. The fix was = to upgrade the FW.  H As a suggestion, given the older type of systems you are using for the =H test, perhaps have CS verify the HBA's you are using have the latest FW?   Regardse  
 Kerry Main Senior Consultants Hewlett-Packard (Canada) Co.! Consulting & Integration Servicest Voice: 613-592-4660n Fax   : 613-591-4477 Email: kerryDOTmain@hpDOTcom   (remove the DOT's)       -----Original Message-----2 From: Craig Cooke [mailto:ccooke@beta.dabs.com]=20 Sent: October 14, 2002 6:40 AM To: Info-VAX@Mvb.Saic.Comb& Subject: Re: VMS 7.3-1 File Corruption     Hi Rener  C No, we do not have any disk defragmentation tools running on this =p system.    Thanks for the input anyway.   Craig Cookeg Dabs.com      A "Ren=E9 Schelbaum" <rene.schelbaum@datakom.at> wrote in message =hA news:3daa9a7c$0$18866$91cee783@newsreader02.highway.telekom.at...  > Hi!: >5H > Do you have a 3rd-party defragmentation-tool (like diskeeper) running? >iH > Had the same problem, because I had forgotten to make sure, that it is still  > compatible - it was not. >p	 > regardst >e > Ren=E9 >P >o@ > "Craig Cooke" <ccooke@beta.dabs.com> schrieb im Newsbeitrag=20! > news:3daa84f8$1@194.70.94.92...o > > Hi,. > >sH > > I've recently upgraded a test system (Alphaserver 1200) - to OpenVMS	 > 7.3-1 -RJ > > storage is via HSG80 - I created a 6 member mirrored stripe set.  We =  H > > had problems with files 'disappearing' / becoming corrupt etc.  I=20
 > > did an > >e > > ANAL/DISK/REP 'DISK' > >UC > > Now, massive corruption has taken place - we get errors such asn > >h< > > %ANALDISK-W-MULTALLOC, file (1,1,0) [000000]INDEXF.SYS;1% > >         multiply allocated blocksl > >         VBN 78490 to 79515H > >         LBN 2088252 to 2089277, RVN 1 %ANALDISK-W-MULTALLOC, file=20  > > (1,1,0) [000000]INDEXF.SYS;1% > >         multiply allocated blocksr > >         VBN 79516 to 144324i > >fC > > %ANALDISK-W-BAD_NAMEORDER, filename ordering incorrect in VBN 3e- > >         of directory 'DIRECTORY' (12,1,1)s% > >         Filenames are 'FILE1'.ISMe% > >                   and 'FILE2.ISM't > > G > > We intend to upgrade out production systems shortly from OpenVMS=20m	 > > 7.2-1n toI > > OpenVMS 7.3-1 (we need to as a new ES45 is being delivered today -=20lH > > and requires commisioning as soon as possible), but obviously, if=207 > > this is a known issue, it requires resolving first.t > >T* > > Does anyone know anything about this ? > > D > > Is it a known issue (I know it was a problem prior to VMS 5.5-2) > >0% > > Thanks ins advance for any input.s > >. > > Craig CookeF > > Dabs.com > >u > >- > >i > >f > >e > >t > >  > >g >D >B   ------------------------------  % Date: Tue, 15 Oct 2002 22:43:36 -0400D- From: JF Mezei <jfmezei.spamnot@videotron.ca>S@ Subject: Re: VMS MAIL: COMPRESS _increases_ the size of the file+ Message-ID: <3DACD253.E1EA671@videotron.ca>:   Phillip Helbig wrote: E > (In case it's not clear, I mean MAIL> COMPRESS, which I think callsi > CONVERT/RECLAIM)..  M CONVERT/RECLAIM does this "in-situ" so there wouldn't need to be an intrinsiceH need for a MAIL.OLD and MAIL.MAI to exist (and remember that at least inE previous versions, MAIL COMPRESS would fail be default because it was L attempting to delete or rename a file that was created by default to preventN the user from deleting it). Now sure how it behaves in recent versions of VMS.  M Now, if it uses and actual CONVERT, perhaps it generates some FDL with targeth4 free space that results in more space being needed.   M I.e. first file was first created with 100 blocks, assuming 70% fill. (eg: 70aN blocks of data) You load the file up with emails which raises the fill rate to7 say 90%,  so the file is still the same allocated size.n  K But when you "Compress" it, it will create a file where 90 blocks representTE 70% of the total size., hence the total file would become 128 blocks.   * The above is just speculation on my part.)  B Looking at sys$library:mailshr.exe , it seems that MAIL calls both CONV$RECLAIM and CONV$CONVERT.( Not sure what part of mail calls either.   ------------------------------    Date: 15 Oct 2002 15:07:45 -0600- From: koehler@encompasserve.org (Bob Koehler) * Subject: Re: Why is LTA0: offline mounted?3 Message-ID: <gK$Dk3$zsizb@eisner.encompasserve.org>   n In article <b096a4ee.0210141204.6cc57f47@posting.google.com>, spamsink2001@yahoo.com (Alan E. Feldman) writes:   > B > Also, is there somewhere that all these odd devices, like EZA0:,G > NOA0:, PKA0:, etc., are documented? I've never run across them in therH > docs (except for NLA0: which is mentioned under SPAWN and maybe one orH > two other random places). And I do know that NLA0: is the null device.  A    I once wrote two drivers for a custom netowrk interface calledeD    NODRIVER and NIDRIVER.  These I loaded to NOA0: and NIA0:.  LaterC    I renamed them to QODRIVER and QIDRIVER when at a national DECUS A    I was told that first letters Q and Z are reserved to customern    device drivers.      ------------------------------    Date: 16 Oct 2002 04:14:43 +0800, From: Paul Repacholi <prep@prep.synonet.com>* Subject: Re: Why is LTA0: offline mounted?- Message-ID: <87n0pfh32k.fsf@prep.synonet.com>o  4 rdeininger@mindspring.com (Robert Deininger) writes:  ? > In article <b096a4ee.0210141204.6cc57f47@posting.google.com>,,1 > spamsink2001@yahoo.com (Alan E. Feldman) wrote:g  sC > >Also, is there somewhere that all these odd devices, like EZA0:,hD > >NOA0:, PKA0:, etc., are documented? I've never run across them inF > >the docs (except for NLA0: which is mentioned under SPAWN and maybeC > >one or two other random places). And I do know that NLA0: is them > >null device.  Thanks again.  -F > I don't know of a manual that documents them all.  You can sometimesF > extract clues by searching SYS$SYSTEM:SYS$CONFIG.DAT for the first 2F > letters of the device name, and looking at the nearby lines of text.  AE Also in the help files under the `system_files' stuff. Some are good,p some arenon-existant :(   ? > Once you know what _kind_ of device it is, the device supporta4 > manuals often detail the supported QIO operations.  rF > EZA0: is an ethernet device.  PKA0: is a SCSI adapter.  It's part of' > the I/O path for DKA and MKA devices.o  iB > Never heard of NOA0:.  Maybe that's the device that NOAA uses to5 > control the weather.  Is it a template device?  :-)    -- e< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda.s@                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Wed, 16 Oct 2002 00:21:45 +0100 4 From: John Laird <john@laird-towers.freeserve.co.uk>* Subject: Re: Why is LTA0: offline mounted?8 Message-ID: <ki8pqust6kv2svj4q219s94l9mi9k5mrej@4ax.com>  E On Mon, 14 Oct 2002 22:28:43 -0400, rdeininger@mindspring.com (Robertn Deininger) wrote:G  I >Never heard of NOA0:.  Maybe that's the device that NOAA uses to controld, >the weather.  Is it a template device?  :-)  L NODRIVER is used if you want to run async Decnet over terminal lines.  DDCMPK and all that.  You manually install it, and then you can dynamically switche lines to network lines.   L I don't personally recall it creating a template device, but then I probably never cared to look.  J Of course, if NODRIVER is unrelated to NOA0:, someone needs their knuckles
 rapped ;-)     	John    ------------------------------  + Date: Tue, 15 Oct 2002 19:28:22 -0500 (CDT)r From: sms@antinode.org* Subject: Re: Why is LTA0: offline mounted?) Message-ID: <02101519282271@antinode.org>-   WUSS $ show device /full no   N Device NOA0:, device type unknown, is online, device set /NOAVAILABLE, network     device..  O     Error count                    0    Operations completed                  0 O     Owner process                 ""    Owner UIC                      [SYSTEM]fO     Owner process ID        00000000    Dev Prot              S:RWPL,O:RWPL,G,W O     Reference count                0    Default buffer size                 256o  E    So, not a template, but still unusual.  In the days before DSL, myeH home VAXsta 2000 used dial-up async DECnet to connect to the VAX 4300 atE work, and then with Wollongong DBRIDGE (IP over DECnet), the Interneto. was just 9600b/s away.  Ah, the good old days.  4 From: John Laird <john@laird-towers.freeserve.co.uk>G > On Mon, 14 Oct 2002 22:28:43 -0400, rdeininger@mindspring.com (Robertt > Deininger) wrote:m > K > >Never heard of NOA0:.  Maybe that's the device that NOAA uses to control0. > >the weather.  Is it a template device?  :-) > N > NODRIVER is used if you want to run async Decnet over terminal lines.  DDCMPM > and all that.  You manually install it, and then you can dynamically switche > lines to network lines.  > N > I don't personally recall it creating a template device, but then I probably > never cared to look.  H ------------------------------------------------------------------------  C    Steven M. Schweda               (+1) 651-699-9818  (voice, home)UC    382 South Warwick Street        (+1) 763-781-0308  (voice, work)2G    Saint Paul  MN  55105-2547      (+1) 763-781-0309  (facsimile, work)j9    sms@antinode.org                sms@provis.com  (work)r   ------------------------------    Date: 15 Oct 2002 18:59:33 -0700. From: spamsink2001@yahoo.com (Alan E. Feldman)< Subject: Re: Why slow system when using logical search list?= Message-ID: <b096a4ee.0210151759.3c8bdd9f@posting.google.com>_  b bleau@umtof.umd.edu (Lawrence Bleau) wrote in message news:<ao2kr5$sg1$1@grapevine.wam.umd.edu>...O > Well, folks, I've done more experimenting, and have more data for you on thise > problem.  A brief recap: > G > When I use a search list to do an open to a file (in a test program),-K > with one of the elements of the search list being a remote system and theMN > others being local disks, and the files actually reside onthe remote system,P > it takes about 45 secs to read through a file, during which time there is highK > kernel mode and interrupt mode activity (35% and 30%).  If I redefine thebN > logical name to not be a search list but point only to the remote system (soA > there is still network traffic), it takes only 8 secs per file.b > O > I wrote another short test program that eliminated all subroutine, simplified L > the test case quite a bit.  I created a variable length file, unformatted,P > 40 byte records, with 10000 records.  I read it using Fortran unformatted READL > of a 10-element integer array.  I open the file with the READONLY keyword. > N > When I ran the test program it again took much more time using a search listK > than without one: 25 secs per file vs. 4 secs.  Also, the processor modesrG > changed greatly: the kernel+interrupt mode time went from 16% to 73%.  > Q > At Hein's suggestion I set FAL$LOG to 7 and looked at the net$server.log file.  Q > This showed something *very* interesting.  To shorten the dump, I decreased thev2 > size of the test files from 10000 to 10 records. > N > In the no search list case, the client system sent *one* GET request and the= > server (remote) system returned 10 records as DAT messages.m > O > In the search list case, the client system sent one GET request at a time andMP > the server system returned one DAT message for each, with a lot of handshakingM > messages in between.  My theory is that this is where the extra kernel modefE > time is being spent, sending those overhead message back and forth.p > P > Does anyone have an idea why Fortran RTL (or RMS) would change its behavior inP > this manner?  Let me know if you want the details of the transactions; I saved > the net$server.log file. >  > Lawrence Bleau > University of Maryland$ > Physics Dept., Space Physics Group > 301-405-6223 > bleau@umtof.umd.edu     B Just some ideas -- no guarantee that any of them will work -- just0 some things to try if other ideas don't pan out:   1.)   - $ DEFINE REMOTE_DIRS  NODE1::DISK1:[DIR1],...u& $ DEFINE LOCAL_DIRS   DISKA:[DIRA],...C $ DEFINE SEARCH_LIST  LOCAL_DIRS:,REMOTE_DIRS:  !! Don't forget the  trailing colons!  A and use SEARCH_LIST. Yeah, I don't see why this would work, but IyE equally don't see why you're having this really strange problem. I'll9$ give it an 8% likelihood of working.  D 2.) Did you try this with DCL and does it have the same problem? Try with TYPE, COPY, and READ.  D 3.) Set the SQO bit for SeQuential access Only (involves setting one? or more of those FAB$x_xxxx, RAB..., XAB things appropriately).e   Disclaimer: JMHO Alan E. Feldmanu   ------------------------------  % Date: Tue, 15 Oct 2002 23:25:04 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>I, Subject: Re: ZDNet says you're dead by 2004!, Message-ID: <3DACDC08.62C00E19@videotron.ca>   Ken Robinson wrote: H > First production release is targeted for early 2004 (version 8.2, this, > version will ship for Alpha and VAX also).  H Shouldn't the initial VMS releases targeted to developpers be comsideredM "betas" (for instance 8.0beta1 beta2 etc) and the first production release bel called 8.0 ?  H Seems strange to be going from 7.3-1 to 8,.2 without a 8.0. I think thatK customers will clue in to the fact that 8.2 is really 8.0 from the point ofM view of reliability.  , Will there be 8.1 and 8.2 on alpha and vax ?   ------------------------------  % Date: Tue, 15 Oct 2002 23:49:54 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>h, Subject: Re: ZDNet says you're dead by 2004!, Message-ID: <3DACE1D8.B1A3CA03@videotron.ca>  ! Someone whose killfiled me wrote:nK > > ...and production-quality, ready-for-prime-time Itanics will be out ...t	 > > when?h > 1 > Before HP stops selling Alphas.  By definition.   3 If ome is strictly speaking, is the above correct ?h  J Isn't the "commitment" for ALPHA purely a number of years (i.e. until soldM until at least 2006 with  support for 5 years after) irrespective of how IA64i does ?  L In other words, assuming a theoretical worse case scenario with Intel unableK to produce enterprise-capable IA64 until 2007, does HP's current commitment L automatically extend Alpha to at least 2007 ? Or is there just an unofficialK expectation that should such a delay happen, HP would extend Alpha's life ?r  N (I realise that it is very likely that HP will declare IA64 "enterprise ready"M well before 2006 so the question is moot, however, strictly speaking I am not4M sure the statement "IA64 will become enterprise ready before HP stops selling $ Alpha. *BY DEFINITION*", is correct.   ------------------------------   End of INFO-VAX 2002.571 ************************madgoat.com/mmk>  > Create a directory, unzip, link, and define "MMK" as a foreign command.  E Sorry not to have a smaller reproducer, but you'll then need to get ar Perl distribution from                                                                                                                                                                                                                                                                