1 INFO-VAX	Thu, 07 Dec 2000	Volume 2000 : Issue 683       Contents:+ Re: "process crash" vs. "application crash" + Re: "process crash" vs. "application crash" + Re: "process crash" vs. "application crash" + Re: "process crash" vs. "application crash" + Re: "process crash" vs. "application crash" + Re: "process crash" vs. "application crash" + Re: "process crash" vs. "application crash" + Re: "process crash" vs. "application crash" , Re: ??== Allocating only one block per file. A great Shockwave flash movie 1 Re: Accessing a VMS system from a SUN workstation 1 Re: Accessing a VMS system from a SUN workstation  Re: Connecting a CD-writer.. Re: Connecting a CD-writer.. Re: Connecting a CD-writer.. Re: DEC BASIC : relative files Re: DEC BASIC : relative files Re: educational licenses eSNMP breaks signal()?8 Has any used the OnStream tape drive on OpenVMS systems? Re: how to sort a CSV file. Re: I need macro32 code for string replacement. Re: I need macro32 code for string replacement IO channels quota # MACRO32 - Oh my God, Iwas so stupid   Re: Need help, thanks in advance Re: New OpenVMS  Education site  Re: New OpenVMS  Education site  Re: Pathworks to NT migration , Re: Problems installing MQ Series on VMS 6.2, Re: Problems installing MQ Series on VMS 6.2 RAID array informations... Re: raid system needed) Re: Source/availability of 60m DAT media? ) RE: Source/availability of 60m DAT media? ) Re: Source/availability of 60m DAT media? % Source/availability of 60m DAT media?  Re: Sun Cluster  Re: Sun Cluster  Re: Sun Cluster  Re: Sun Cluster  Re: Sun Cluster  Re: Sun Cluster  Re: Sun Cluster  Re: Sun Cluster  Re: Sun Cluster ) Sun CSV Constitution Election Belly Laugh - Re: Sun CSV Constitution Election Belly Laugh I Re: Sun new cluster comparison ommitted VMS (but VMS was their benchmark) I Re: Sun new cluster comparison ommitted VMS (but VMS was their benchmark)  Testers wanted for Spell update ; Trying to get Sevice and Support for a Couple of VMS Alphas  Re: US Constitution  Re: Very weird DCL behavior  Re: Very weird DCL behavior  Re: Very weird DCL behavior  RE: Very weird DCL behavior ( Why do I have to run DECW$UTILS:XREFRESH XP1000 hardware questions  Re: XP1000 hardware questions   F ----------------------------------------------------------------------    Date: 07 Dec 2000 11:25:52 +0100G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de> 4 Subject: Re: "process crash" vs. "application crash"H Message-ID: <y4d7f4h6j3.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  N As I wrote before, in all likelihood you're triggering an RMS bug - it's a bugK by definition because it's killing your process. Set BUGCHECKFATAL to 1 and N wait for the system crash that will be triggered by this to occur. Be sure youL can tolerate that, and that the system dump file is set up correctly so that you can analyse it properly.    ' You can also start off by checking with J http://ftp.support.compaq.com/patches/.new/openvms.shtml that you have all* relevant patches installed on your system.   	Jan   ------------------------------  % Date: Thu, 07 Dec 2000 08:24:10 -0500 , From: Howard S Shubs <hshubs@mindspring.com>4 Subject: Re: "process crash" vs. "application crash"> Message-ID: <hshubs-3C1410.08241007122000@news.mindspring.com>  J In article <90n6i8$l3n$1@nnrp1.deja.com>, david_dawkins@my-deja.com wrote:   >Hallelujah!!!  O RMS runs in executive mode, you see.  If the process bombs while in exec mode,  A it's already too late to give control to RMS, as I understand it.     A >Right, in effect, I'm the developer; I'm certain none of my code @ >is going into executive mode (I don't even know how to do that,= >or why I would want to); however we do link with third-party = >libraries for which we have the source. Can you tell me what = >syscall(s) I should be looking for, or what else to look for / >in those libraries to identify exec-mode code?   " $CMEXEC (Change Mode to EXECutive) --   Howard S ShubsD "Run in circles, scream and shout!"  "I hope you have good backups!"   ------------------------------   Date: 7 Dec 2000 08:48:17 -0500 9 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) 4 Subject: Re: "process crash" vs. "application crash"+ Message-ID: <TZW6BCO++hs1@eisner.decus.org>   K In article <90n1u4$hgu$1@nnrp1.deja.com>, david_dawkins@my-deja.com writes:   > > so I have this detached process that dies and never leaves a6 > dump file; quota and resource limit issues have been* > eliminated by people more clever than I.  / What makes you expect it to leave a dump file ?   . Have you started it with the /DUMP qualifier ?  K What is the process termination code from Accounting ?  That always exists.   = > - Could this type of crash explain the consistent lack of a  > stack trace and dump files?   . Yes, but so could a whole lot of other causes.  @ > - what can we do to prepare a process such that we can collect@ > diagnostics when such a process crash occurs (such as a dump)?  I If you set system parameter BUGCHECKFATAL what you call a "Process Crash" : will crash the whole system instead, creating a dump file.  N ==============================================================================N Great Inventors of our time: Al Gore -> Internet; Sun Microsystems -> ClustersN ==============================================================================   ------------------------------  # Date: Thu, 07 Dec 2000 14:07:58 GMT  From: david_dawkins@my-deja.com 4 Subject: Re: "process crash" vs. "application crash") Message-ID: <90o5jq$c63$1@nnrp1.deja.com>    Thanks Jan,   D the system guys havealready tried that; no luck. I should have said.   Thanks,    David   H In article <y4d7f4h6j3.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>,C   Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>  wrote:E > As I wrote before, in all likelihood you're triggering an RMS bug - 
 it's a bugG > by definition because it's killing your process. Set BUGCHECKFATAL to  1 and G > wait for the system crash that will be triggered by this to occur. Be  sure youF > can tolerate that, and that the system dump file is set up correctly so that  > you can analyse it properly. > ) > You can also start off by checking with C > http://ftp.support.compaq.com/patches/.new/openvms.shtml that you  have all, > relevant patches installed on your system. >  > 	Jan >     & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  # Date: Thu, 07 Dec 2000 14:32:37 GMT  From: david_dawkins@my-deja.com 4 Subject: Re: "process crash" vs. "application crash") Message-ID: <90o726$dg2$1@nnrp1.deja.com>   + In article <TZW6BCO++hs1@eisner.decus.org>, <   Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) wrote:E > In article <90n1u4$hgu$1@nnrp1.deja.com>, david_dawkins@my-deja.com  writes:  > @ > > so I have this detached process that dies and never leaves a8 > > dump file; quota and resource limit issues have been, > > eliminated by people more clever than I. > 1 > What makes you expect it to leave a dump file ?  > 0 > Have you started it with the /DUMP qualifier ?  7 yes; we also rigged the program to crash with an ACCVIO @ (by evaluating *(int*)1) when sent a particular message "crash".  > > What is the process termination code from Accounting ?  That > always exists.  A I asked the system guys, they say they can never find it. They've ( been tracking this issue for a year now.    ? > > - Could this type of crash explain the consistent lack of a  > > stack trace and dump files?  > 0 > Yes, but so could a whole lot of other causes.  < In understand.  Can you name some?  Bear in mind that we can& get dump files when we fake the crash.  B > > - what can we do to prepare a process such that we can collectB > > diagnostics when such a process crash occurs (such as a dump)? > 9 > If you set system parameter BUGCHECKFATAL what you call A > a "Process Crash" will crash the whole system instead, creating  > a dump file.  9 OK, this has been done already, and no dump file was ever ? produced (no system crash).  Does this mean we definitely don't  have a "process crash"?   9 I'm keen to get the terminology right; what do you call a  "process crash"?   Thanks,   
 David Dawkins     & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------    Date: 07 Dec 2000 15:37:42 +0100G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de> 4 Subject: Re: "process crash" vs. "application crash"H Message-ID: <y4aea8i9ft.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  ! david_dawkins@my-deja.com writes:   F > the system guys havealready tried that; no luck. I should have said.   That would be unprecedented.  K An exception in user mode won't kill the process, and you will at least get L the dump from the last chance handler. An exception in supervisor mode (DCL)I _will_ kill the process, but you will get the dump from DCL's last chance L handler (use a paper terminal if possible, or an xterm or similar with largeE scroll buffer) - I have seen those happen when hardware was dying. An J exception in exec mode (RMS) _will_ kill the process without notice to theM user, but BUGCHECKFATAL=1 will crash the system in that case. An exception in F kernel mode will always crash the system unless the code producing the  exception has taken precautions.    What about the accounting entry?   	Jan   ------------------------------  # Date: Thu, 07 Dec 2000 15:06:03 GMT  From: david_dawkins@my-deja.com 4 Subject: Re: "process crash" vs. "application crash") Message-ID: <90o90m$f6n$1@nnrp1.deja.com>   9 > > >  [ set bugcheck=fatal to catch an exec-mode crash ] 7 > > the system guys have already tried that; no luck. I  > > should have said.  >  > That would be unprecedented.  4 Hmmm... so it's not looking like an exec-mode crash.  " > What about the accounting entry?  D I am told that there is no accounting entry for the process. Is this all sounding too weird?  --9 So I have no accounting info, no stack trace, no dump, no 7 crash with bugcheck=fatal, (and no clue).  what's left?    David     & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  % Date: Thu, 07 Dec 2000 15:19:40 +0000 - From: Tim Llewellyn <tim.llewellyn@bbc.co.uk> 4 Subject: Re: "process crash" vs. "application crash") Message-ID: <3A2FAA8C.AF77658A@bbc.co.uk>    Jan Vorbrueggen wrote:  # >  What about the accounting entry?  >   ? Yup, if you enable process termination accounting you should at ; least have a completion status coda in there to start from.    > 
 >         Jan    --6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uk   A I speak for myself only and my views in no way represent those of  MedAS or the BBC.    ------------------------------  % Date: Thu, 07 Dec 2000 08:20:24 -0500 , From: Howard S Shubs <hshubs@mindspring.com>5 Subject: Re: ??== Allocating only one block per file. > Message-ID: <hshubs-B953D7.08202407122000@news.mindspring.com>  E In article <aus-0712000946450001@wvia26.virologie.uni-wuerzburg.de>,  - aus@vim.uni-wuerzburg.de (Hans M. Aus) wrote:   I >Is it possible to allocate only one block per disk file without changing  >the cluster size on the disk?  / Sure, if the cluster size is already one block.     J >Is it possible define a default file allocation to just one block for all( >those files received by the FTP server?  M Sounds like you don't understand what "cluster size" means.  Cluster size is  H the minimum number of blocks allocatable.  You can't get fewer than the 8 minimum, and this minimum is set at initialization time.  O Here in the Macintosh world, HFS disks have a limitation of 65536 clusters per  N volume.  Therefore, the larger the volume, the bigger the cluster size.  It's M similar under VMS, except that VMS doesn't have that particular limitation.   M Under VMS, I recently set up certain disks with a cluster size of 2048.  I'm  M not sure my concept of "easier to allocate big files" was successful, but it  O sounded okay.  Note: the disk block size under VMS and with a Macintosh is the   same: 512 bytes.    I >What are the negative consequences of initializing one disk to a cluster  >size of one block?   L It'd become "harder" to deal with larger files.  INDEXF.SYS would be large, N the allocation bitmap would be large, fragmentation would be more of an issue.    3 >Is there a reasonable compromise such as 3 blocks?   O Three blocks is good, unless you need the additional disk space you'd get with  O an cluster size of one block.  One way to be sure you'll get a smaller cluster  M size (by default) is to use a smaller disk.  The cluster size of a device is  ) calculated based on the size of the disk.   K I've set Newsgroups and Followups appropriately, as this isn't a Macintosh   question but a VMS one.  --   Howard S ShubsD "Run in circles, scream and shout!"  "I hope you have good backups!"   ------------------------------  $ Date: Wed, 6 Dec 2000 13:52:02 -0500+ From: "Ari, Nasit -AES" <nasit.ari@itt.com> & Subject: A great Shockwave flash movie] Message-ID: <F84F1E97B3FAD31186FA009027DCBE29B2298F@reston-exchange.reston.aes.de.ittind.com>   H Check out this new flash movie that I downloaded just now ... It's Great Bye    ------------------------------  $ Date: Tue, 5 Dec 2000 22:16:24 +0100. From: "Marc Van Dyck" <marc.vandyck@skynet.be>: Subject: Re: Accessing a VMS system from a SUN workstation* Message-ID: <90jlsl$b17$1@news1.skynet.be>   Thanks to all who have replied.   L With your help, my Unix specialist has been able to solve the problem, using the following method :  K 1 an input file to Xmodmap to re-assign the keys so that they look as much  asI      possible like a VMS keyboard. This also disables the local effect of  the 'num lock'I      key, so that it can act as the usual gold key on the remote machine;   K 2 a second input file to Xmodmap to reset the normal situation, which also  re- #      establishes the num lock key ;   H 3 a toggle program, that activates either 1 or 2, driven by two keys, declared as /      hot keys in the setup of my motif session.   B I'm at home so I can not give much more details, but if someone isI interested, I can bring the files from the office on a floppy and publish 
 them here.  $ Now I have something that I can use.   Marc.   L "Martin Vorlaender" <martin@radiogaga.harz.de> a crit dans le message news:0 3a248310.524144494f47414741@radiogaga.harz.de.../ > Marc Van Dyck (marc.vandyck@skynet.be) wrote: H > : So, the question : does anyone know the right shell incantation, the right K > : file (X resource ?) to modify, to make that keypad usable ? I have some  UnixG > : experts available on site, so if a solution is suggested, I imagine  they'll + > : be able to understand and implement it.  >  > Hope this helps,
 >   Martin >   (who uses eXceed)  >  > [-- snip --] > From: oberman@icaen.llnl.gov > Newsgroups: vmsnet.tpu6 > Subject: RE: Wanted: EDT/TPU keypad bindings for SUN# > Date: Thu, 11 May 95 19:03:29 GMT 6 > Organization: Lawrence Livermore National Laboratory- > Message-ID: <3otjg6$rf@lll-winken.llnl.gov>  > < > larsen@beethoven.sedd.trw.com (Eric Hadley Larsen) writes:K > > I have a Sun Sparc on my desktop.  I often remotely create a DECterm on H > > my VAX/VMS machine and route the display back to by sun.  This works great K > > until I try to edit a file.  I cannot seem to figure out to get the EDT 6 > > style keypad to map correctly for my SUN keyboard. > G > This can get VERY complex. But I can give you a start. Use xmodmap to & > remap your keys to do what you want. > J > You don't say what type of Sun keyboard you have. I have the Type 4 thatI > lacks any cursor keys except on the application keypad, so I think it's  > about the worst case.  > : > Here is the file I pass to xmodmap in my .xsession file: >  > keycode 101=KP_0 > keycode 119=KP_1 > keycode 120=KP_2 > keycode 121=KP_3 > keycode 98=KP_4  > keycode 99=KP_5  > keycode 100=KP_6 > keycode 75=KP_7  > keycode 76=KP_8  > keycode 77=KP_9  > keycode 28=KP_F1 > keycode 29=KP_F2 > keycode 30=KP_F3 > keycode 57=KP_DecimalX > keycode 52=Prior > keycode 53=Next8 > keycode 54=Find  > keycode 105=KP_F4s > keycode 78=KP_Subtract > keycode 8=Left > keycode 10=Right > keycode 32=Ups > keycode 33=Downa > keycode 132=KP_Separator > keycode 58=Menui > keycode 97=KP_Enters >aF > This maps the applications keypad to do the "normal things, with theF > extra keys in the second row from the top mapped to prior, next, andI > find. The top 4 "L" keys are mapped to left, right, up, and down. L6 isa > the "DO" key.  >iJ > In addition, there are odd Motif related problems with several keys that8 > may be taken care of with an X resource in .Xdefaults: >  > *defaultVirtualBindings:\ 5 >  osfCancel      :               <Key>F11        \n\f5 >  osfLeft        :               <Key>Left       \n\i5 >  osfUp          :               <Key>Up         \n\a5 >  osfRight       :               <Key>Right      \n\ 5 >  osfDown        :               <Key>Down       \n\e5 >  osfEndLine     :Alt            <Key>Right      \n\l5 >  osfBeginLine   :Alt            <Key>Left       \n\d5 >  osfPageUp      :               <Key>Prior      \n\t5 >  osfPageDown    :               <Key>Next       \n\N5 >  osfDelete      :Shift          <Key>Delete     \n\65 >  osfUndo        :Alt            <Key>Delete     \n\a5 >  osfBackSpace   :               <Key>Delete     \n\e5 >  osfAddMode     :Shift          <Key>F8         \n\t5 >  osfHelp        :               <Key>Help       \n\T5 >  osfMenu        :               <Key>F4         \n\u5 >  osfMenuBar     :               <Key>F10        \n\ 5 >  osfSelect      :               <Key>Select     \n\l5 >  osfActivate    :               <Key>KP_Enter   \n\s5 >  osfCopy        :Shift          <Key>DRemove    \n\n5 >  osfCut         :               <Key>DRemove    \n\o- >  osfPaste       :               <Key>Insert  > I > Note that this must be EXACTLY right or nothing will work, so you might  > want to cut and paste it.i >tF > now if I could jsut come up with a Type 5 keyboard I could eliminateI > most of this and go faster because I'd have the keys where they belong.i >D > R. Kevin Oberman! > Energy Sciences Network (ESnet)v7 > National Energy Research Supercomputer Center (NERSC) / > Lawrence Livermore National Laboratory (LLNL)-- > Internet: koberman@llnl.gov +1 510-422-6955- >- > --L > One OS to rule them all       | Martin Vorlaender  |  VMS & WNT programmer9 > One OS to find them           | work: mv@pdv-systeme.dev! > One OS to bring them all      | ( http://www.pdv-systeme.de/users/martinv/@ > And in the Darkness bind them.| home: martin@radiogaga.harz.de   ------------------------------  % Date: Thu, 07 Dec 2000 08:28:35 -0500   From: jamese@beast.dtsw.army.mil: Subject: Re: Accessing a VMS system from a SUN workstation0 Message-ID: <00120708283579@beast.dtsw.army.mil>  1 "Marc Van Dyck" <marc.vandyck@skynet.be> wrote inoA <90jlsl$b17$1@news1.skynet.be> on Tue, 5 Dec 2000 22:16:24 +0100:w  ! > Thanks to all who have replied.a > N > With your help, my Unix specialist has been able to solve the problem, using > the following method : > M > 10 an input file to Xmodmap to re-assign the keys so that they look as much<L >    as possible like a VMS keyboard. This also disables the local effect ofH >    the 'num lock' key, so that it can act as the usual gold key on the >    remote machine; > M > 20 a second input file to Xmodmap to reset the normal situation, which alsot& >    re-establishes the num lock key ; > J > 30 a toggle program, that activates either 10 or 20, driven by two keys,; >    declared as hot keys in the setup of my motif session.t > D > I'm at home so I can not give much more details, but if someone isK > interested, I can bring the files from the office on a floppy and publish  > them here.  J Yes, I'm very interested. Please publish your files to info-vax. Thank you and thanks to your Unix guru.o  : Ed James                           ed.james@telecomsys.com5 TeleCommunications Systems, Inc.   voice 410-295-1919n; 2024 West Street, Suite 300              800-810-0827 x1919s5 Annapolis, MD 21401-3556           fax   410-280-10940   ------------------------------  $ Date: Thu, 7 Dec 2000 13:33:15 +01001 From: "Willem Grooters" <Willem.Grooters@sema.nl>r% Subject: Re: Connecting a CD-writer..d' Message-ID: <UoLX5.1913$104.923@client>r   > >> >J > >> > We don't have the luxury of making customizations to DKDRIVER in anH > >> > effort to make some less retrictive compromises than were made by OVMS > >>C > >> You can write your own DKDRIVER and implement it, if you wish.tA > >> I was involved in a project that did just that a while back.i > >?1 > > Have you any code you can post or contribute?n > >? >h= > Not my code, and for a company I no longer work for, so thec: > answer would be no.  Besides, I wasn't the brains of the> > operation.  At the time were actually trying to get around a< > known bug in the DK driver that was causing the product in@ > question to not see all the I/Os going through it.  This was a/ > long time ago in a VMS version far, far away.o >s< > Glen Everhart (spelling?) has done a number of examples, IA > thought.  Aren't his freeware disk bits just variations on that=C > theme?   Hasn't he made significant contributions in the DKDRIVERtA > stuff for VMS in the past?  Maybe he could provide the freeware=D > folks with an example DKlike driver example, if he hasn't already? >-B > Schekenberger could probably knock one out during the commercial? > breaks in the X-files.  He V7-ified, fast-pathed and found aneC > uncovered bug in a compression disk driver product I used supportm= > in about that time.  It looked to VMS like a plain old diskhB > (except for the ZR device prefix).  Maybe he, too, could provide; > the freeware folks with an example DKlike driver example?r  L I missed the original question and the following answers - I just subscribedD to this group - but I'm anxious to know how to do this (Compaq DS20E system).= We don't have this luxury to write a driver ourselves either.sJ So could I be updated : Which drives are supported? What software needs to) be installed and/or configured (and how)?i   Thanks in advance.   Willem Grootersn Willem.Grooters@sema.nl    ------------------------------  # Date: Thu, 07 Dec 2000 13:23:32 GMT/= From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-)h% Subject: Re: Connecting a CD-writer..n0 Message-ID: <009F43AF.732F13A7@SendSpamHere.ORG>  X In article <q+IVpxg0nX4U@eisner.decus.org>, kuhrt@eisner.decus.org (Marty Kuhrt) writes:h >In article <3A2D88DC.68E8379B@earthlink.net>, "David J. Dachtera" <djesys.nospam@earthlink.net> writes: >> Marty Kuhrt wrote:  {...snip...}A >Schekenberger could probably knock one out during the commercials> >breaks in the X-files.  He V7-ified, fast-pathed and found an  = Gave up on the X-files...  Got boring and excessively stupid.   B >uncovered bug in a compression disk driver product I used support< >in about that time.  It looked to VMS like a plain old disk  7 ... and more recently, updating it for the use og GBOs.e  A >(except for the ZR device prefix).  Maybe he, too, could provider: >the freeware folks with an example DKlike driver example?  < What's the gripe with DKDRIVER?  The problem is the plethora< of third-party SCSI devices and the plethora of their inter- pretations of the SCSI specs.    --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COMs             O city, n., 1. a place where trees are cut down and streets are named after them.t   ------------------------------  # Date: Thu, 07 Dec 2000 13:34:19 GMTc= From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-) % Subject: Re: Connecting a CD-writer.. 0 Message-ID: <009F43B0.F489CA49@SendSpamHere.ORG>  [ In article <UoLX5.1913$104.923@client>, "Willem Grooters" <Willem.Grooters@sema.nl> writes:  {...snip...}M >I missed the original question and the following answers - I just subscribedhE >to this group - but I'm anxious to know how to do this (Compaq DS20Ee	 >system).a> >We don't have this luxury to write a driver ourselves either.K >So could I be updated : Which drives are supported? What software needs to * >be installed and/or configured (and how)?  M Many of the SCSI CDR/RW drives will work just fine.  I have a Yamaha CRW4416SdM which works fairly well.  I've also had the chance to connect a Plextor and a0$ Smart&Friendly device without issue.  M You'll need software to control the CDR/CDRW device for burning your own CDs."L I have links to this software on a page at my web site devoted to CD writing$ on VMS: http://www.tmesis.com/CDrom/   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COMe            tO city, n., 1. a place where trees are cut down and streets are named after them.    ------------------------------  % Date: Thu, 07 Dec 2000 05:40:49 -0800 , From: Jack Patteeuw <jjpatteeuw@voyager.net>' Subject: Re: DEC BASIC : relative files 8 Message-ID: <3a2f68c5$0$35015$2c3edae7@news.voyager.net>   hein@eps.zko.dec.c*m wrote:e  J > Sure... check out an answer I gave to a similar question 10 years ago...    " Thank God old VMSers never die !!!  
 Jack Patteeuwn   ------------------------------  % Date: Thu, 07 Dec 2000 12:31:02 +0100   From: Paul Sture <paul@sture.ch>' Subject: Re: DEC BASIC : relative filesp+ Message-ID: <VA.000001bb.0df22097@sture.ch>.  H In article <KrzNbgAwwqL6EwqI@wrightnet.demon.co.uk>, Steve Wright wrote:3 > From: Steve Wright <usenet@wrightnet.demon.co.uk>i > Newsgroups: comp.os.vmsj) > Subject: Re: DEC BASIC : relative filesr& > Date: Wed, 6 Dec 2000 21:14:24 +0000 > J > In article <3a2d8ef1$2$lllp186$mr2ice@news.flashcom.com>, yyyc186.illega > ltospam_@flashcom.net writes< > >In <FmGeIlAn+XL6Ew6L@wrightnet.demon.co.uk>, on 12/05/00 D > >   at 11:52 PM, Steve Wright <usenet@wrightnet.demon.co.uk> said: > > M > >Stupid question?  Why don't you use an RMS Indexed file?  Simply declare a N > >field or series of fields on your existing record layout as a primary key. K > >You can then do keyed hits for specific records, delete records, and add - > >records simply by using the PUT statement.  > >/I > I tried not to get bogged down into the why we wanted to use a relative I > file in this manner as that involves trying to explain what the program.I > is actually doing. I just wanted to concentrate on the technical aspect2 > of is it possible. >  > However here goes; > J > We have a number of client processes writing to a single server process.B > The primary method of communication is via a mailbox. But as theE > messages are critical and must not be lost in the case of a mailboxyF > failure (server process crash for example) we use a file as a backup > method of communication. > J > Effectively this file is being treated as a FIFO queue. When the clientsH > processes fail to write to the mailbox they open (or create) this FIFOH > file and append their message to the end. They then close the channel. > H > When the server process come back on line, it detects the existence ofG > the FIFO file. It opens the file with ALLOW NONE clause and processesnF > the records. As each record is processed it is deleted from the file@ > just in case the server process crashes again part way throughE > processing the file. Before closing the channel, the server processnF > issues a KILL to delete the file. This prevents any client processesF > from appending records to a file the server thinks it has processed. > J > We know that we cannot delete records from a sequential file (apart fromB > SCRATCH to the end which is of no use here), but we could deleteJ > individual records from a relative file. The only problem we had was howI > to write a record to the end of the file. We were not bothered what thetJ > record number was, nor did we want to have the overhead of maintaining aF > control record we simply wanted to append a record to the end of the > file.  > J > It would be nice if BASIC allowed the use of ACCESS APPEND with relativeH > files so that PUT without a specific record number would append to the" > end of the file. But it doesn't. > I > The problem with using an indexed file is that you need to define a keydG > that is chronological if you are going to use it as a FIFO queue. The C > time available to BASIC is only available in seconds which is notfD > sufficiently small enough. This means we need to delve into systemF > routines in order to use quadword time format. We were looking for a > similar approach.a > O Going back many years I can recall various folks trying to migrate FIFO queues  N onto VMS and using indexed files. This resulted in continuously growing files H as the deleted space was not reclaimed (this before CONVERT/RECLAIM was H available), with a corresponding performance hit. Just another point to 	 consider.b   [rest snipped] ___ 
 Paul Sture Switzerlando   ------------------------------   Date: 6 Dec 2000 09:07:16 GMTi* From: helbig@astro.rug.nl (Phillip Helbig)! Subject: Re: educational licenses". Message-ID: <90kvk4$38e$1@info.service.rug.nl>  H In article <OF41AF6716.8C0ADC13-ON802569AC.00449100@qedi.quintiles.com>,# steven.reece@quintiles.com writes: A  U > http://www.openvms.compaq.com/ebusiness_without_compromise/educational_license.htmle  D > In article <8urhbh$3bm$2@info.service.rug.nl>, helbig@astro.rug.nl > (Phillip Helbig) writes:  H > > re: Discussion of the release of free single-user OpenVMS license vs > > the+< > > (desired) release of free multi-user OpenVMS licenses -- > > licenses for > > use at educational sites.n  E > Would you be kind enough to point me at some info about this.  I'vep: > obviously missed this thread as I've not heard of it.<<<  G Basically, folks are happy that this programme exists.  However, peoplem? in the newsgroup can't understand why there is a restriction topD SINGLE-USER licenses, and no-one from Compaq has commented on this. F True, multi-user licenses exist with DECcampus, but if one already hasC DECcampus, then the new programme is redundant with the significantnD exception that it includes base licenses.  However, many sites don't@ have DECcampus.  It seems that here is potential for expansion. F However, especially at a site where there is no DECcampus, there won'tF be much VMS.  Giving people an account on one's machine is a great wayF to get VMS known.  But this is not possible since all the licenses areH SINGLE-USER licenses.  The only reason I can think of is that this mightD be an incentive to buy more hardware.  However, people interested inD free licenses are not going to buy extra hardware just because of an1 intentional restriction to single-user licenses. s  E Compaq, take away the single-user restriction and non-VMS sites will eC have the potential to get people interested in VMS.  Most of these  9 people will leave academia and BUY hardware and software.g   ------------------------------   Date: 7 Dec 2000 14:30:28 GMTn- From: lewis@lumina.mitre.org (Keith A. Lewis)  Subject: eSNMP breaks signal()?p( Message-ID: <90o6u4$fso$1@top.mitre.org>  J I'm developing an eSNMP subagent under OpenVMS/Alpha 7.2-1, Digital tcp/ipH V5.0A, and Compaq C V6.2-008.  Being under development, it's buggy and IH change it a lot.  It would be really handy to be able to reset it with aK kill() call from another process.  But evidently signal() doesn't work whenS! TCPIP$ESNMP_SHR.EXE is linked in.l  ) I boiled it down to a short test program:e  0 /* SIGL.C Keith Lewis  6-DEC-2000 22:22:36.27 */   /* a signal listener */e   #include <signal>  #include <stdio> #include <stdlib>  #include <unistd.h>    void handler (int param) {$     printf("Param is %d.\n", param);     exit(1); }    main() {     signal(SIGTERM, &handler);     while (1) {.         sleep(5);a         puts("No signal yet.");a     }a }t  	 $ CC SIGLt $ LINK SIGL	 Everything works fine.  	 $ CC SIGL  $ LINK SIGL,SYS$INPUT:/OPT# SYS$SHARE:TCPIP$ESNMP_SHR.EXE/SHARE  $! It never gets the SIGTERM!    : The utility sending the SIGTERM is the same in both cases.  J Can anybody tell me where I went wrong, where Compaq went wrong, or how to+ get around this?  I'd appreciate some help.c  2 --Keith Lewis              klewis@mitre.nospam.org PGP key available.          > The above may not (yet) represent the opinions of my employer.   ------------------------------  % Date: Wed, 06 Dec 2000 22:48:10 -0800a! From: Koloth <koloth@tmisnet.com>RA Subject: Has any used the OnStream tape drive on OpenVMS systems?t* Message-ID: <3A2F32AA.96CE620@tmisnet.com>  B I have a hobbiest OpenVMS system on a personal workstation.  I wasH looking for a tape backup drive.  I was looking at the OnStream 30 GB orG 50 GB tape drive.  Has anyone ever used it on an OpenVMS system before?n  6 I asked and they don't support the fast-skip function.   TIAf   Cass Witkowski   ------------------------------  % Date: Thu, 07 Dec 2000 05:29:51 -0800u, From: Jack Patteeuw <jjpatteeuw@voyager.net># Subject: Re: how to sort a CSV file 8 Message-ID: <3a2f6634$0$35015$2c3edae7@news.voyager.net>   Hoff Hoffman wrote:  > Y > In article <90j7o2$bob$1@nnrp1.deja.com>, Bru, Pierre <Pierre.Bru@spotimage.fr> writes: D > :is there a way to sort a Comma Separated Values file (well, maybeH > :another separator) with the standart sort utlity ? (of course, fields > :are not fixed length...)o > E >    Use some simple DCL to make the records fixed-length, then SORT?     . Sorry Hoff, the problem is "uglier" than that.  F Basically you would have to do a 2 pass parse on the file.  First passD you would have to calculate the maximum size of each field.  In the C second pass, you can then rewrite the fields, padding them out.  OfpF course if it is a numeric field, you probably want it right justified,B and then of course you have to tell SORT that, and if it is a date2 field ... I don't even want to think about it !!!!  
 Jack Patteeuwn   ------------------------------  % Date: Thu, 07 Dec 2000 08:57:46 +0000  From: Roy Omond <Roy@Omond.net> 7 Subject: Re: I need macro32 code for string replacementn) Message-ID: <3A2F5109.C24D7883@Omond.net>I  " "Gotfryd Smolik, VMS lists" wrote:  4 > On 5 Dec 2000 hoffman@xdelta.zko.dec.nospam wrote: > [...]n3 > +  And why Macro32?  Why not some other language?a >p@ >  Because MACRO and DCL the only supported and "additional cost > free" languages ?w >e  + Hey, don't forget TPU which is also "free".y  	 Roy Omondw Blue Bubble Ltd.   ------------------------------  % Date: Thu, 07 Dec 2000 17:24:39 +0800N- From: David B Sneddon <dbsneddon@bigpond.com>n7 Subject: Re: I need macro32 code for string replacemente? Message-ID: <3.0.6.32.20001207172439.00795470@mail.bigpond.com>S  + At 08:57 AM 12/7/00 +0000, Roy Omond wrote:d# >"Gotfryd Smolik, VMS lists" wrote:  >t5 >> On 5 Dec 2000 hoffman@xdelta.zko.dec.nospam wrote:r >> [...]4 >> +  And why Macro32?  Why not some other language? >>A >>  Because MACRO and DCL the only supported and "additional coste >> free" languages ? >> >c, >Hey, don't forget TPU which is also "free". >i
 >Roy Omond >Blue Bubble Ltd.@   ...and so is TECO      Regards, Dave.rI -------------------------------------------------------------------------cI David B Sneddon (dbs)  OpenVMS Systems Programmer   dbsneddon@bigpond.comuI Sneddo's quick guide...           http://www.users.bigpond.com/dbsneddon/ I DBS freeware at ...   http://www.users.bigpond.com/dbsneddon/software.htm I "Life is what happens to you while you're busy making other plans" Lennon    ------------------------------   Date: 7 Dec 2000 17:47:23 GMTa7 From: Thomas.Hahnemann@nospam_s-t.de (Thomas Hahnemann)  Subject: IO channels quota0 Message-ID: <Oozvf8elmJpy-pn2-MdWYxcbf3rKY@Tom2>  ( Which quota limits the maximum number of& IO channels I can assign to mailboxes,0 I've tried to increase fillm but has no success.+ after opening about 210 mb devices I've gots 436, no IO channel available.c     Thanks in advancee   Thomas Hahnemann  s   ------------------------------  $ Date: Thu, 7 Dec 2000 15:52:34 +0100% From: "IdrEASY" <IdrEASY@bigfoot.com>e, Subject: MACRO32 - Oh my God, Iwas so stupid( Message-ID: <90o887$7d89$1@as121.tel.hr>  J Thanx to all people who answer my question about character conversion over MACRO32.  K I made big mistake, because not mask all bits which are needed to translate0 one character to other.M  J Today the lamp in my head was lighten, and this is all because of you good people.h   Here is a little summary:   J Problem with Croatian alphabet and old ASCII table are because we have forF example letter between C and D and this character was substituted with character ^.  L When I sort, string with first character ^  it coming after those which have first character Z.  J We have four more characters like this and "two" which have two characters	 in "one".eI That mean we write them like two characters and read them like one voice. / Lucky thing is that those two are sorting good.   L Now I mask all characters in string so it's parts are characters bigger than ASCII 127 and put inL temporary file indexed file or array together with record number of original characters.lC Then I get this temporary datas sequentially and over record numbern accessing to originals.i' So now I have listing with right order.i  % Sorry because my bad English grammar.v   Bye!   ------------------------------  % Date: Thu, 07 Dec 2000 15:47:39 +0010e% From: paddy.o'brien@zzz.tg.nsw.gov.aui) Subject: Re: Need help, thanks in advance 5 Message-ID: <01JXFFCQZG3M007IEA@tgmail.tg.nsw.gov.au>E   David Dachtera wrote:e >"Philip J. Lewis" wrote:i >> e< >> Legacy is what Unix people say about VMS (and all IBMs op >> ,) >> but they never did know how to count !bL >> Legacy is what MS people say about ALL other opersating systems, but they >> never could spell.d >>  M >> It is always derogatory when used in the systems sense, the implication istG >> that one has inherited these systems (a legacy) and is now forced toRI >> maintain them while newer and, according to the writer/speaker/pointy n haired0 >> one, ipso facto better options are available. >> iH >> The word is also used in database and other IT technology areas.  TheL >> meaning is the same, it is always derogatory and in ALMOST ALL instances  it >> is not accurate.s >>  = >> I would humbly suggest you drop it from your IT vocabularyh >sI >Just my opinion, and perhaps it shows my age (XLVI), but I remember whendG >a legacy was like - make that was - your inheritance: the fortune left-H >to you by your fore-bearers, something to be proud of, a well-spring ofH >value, a treasure, a wealth of knowledge, everything that was built and* >worked for by those who went before us... >a? >Have I really outlived the world which brought me to this day?i  K True, and it gets worse.  Another word that has the same cuddly feeling of nF goodness which I have started to hear being used as a synonym for the J derogatory use of "legacy", is "heritage".  Luckily neither word has this  meaning listed in M-W.   Regards, Paddy   Paddy O'Brien, Transmission Development, 
 TransGrid, PO Box A1000, Sydney South,  NSW 2000, Australia    Tel:   +61 2 9284-3063 Fax:   +61 2 9284-3050& Email: paddy.o'brien@zzz.tg.nsw.gov.au  M Either "\'" or "\s" (to escape the apostrophe) seems to work for most people,r; but that little whizz-bang apostrophe gives me little spam.y   ------------------------------   Date: 6 Dec 2000 15:58:27 GMTi* From: helbig@astro.rug.nl (Phillip Helbig)( Subject: Re: New OpenVMS  Education site. Message-ID: <90lnn3$edl$1@info.service.rug.nl>  ? In article <87n1ear0md.fsf@k9.prep.synonet.com>, Paul RepacholiP# <prep@k9.prep.synonet.com> writes: V  5 > nothome@spammers.are.scum (Malcolm Dunnett) writes:a >  > .... > Q > >    I'm sorry to sound so cynical, but I'm really disappointed in this program M > > After waiting almost a year from the first suggestions it was coming it'snO > > disheartening to see that it really doesn't do much for an institution. TheBQ > > only thing wrong with the CSLG was that it doesn't include base licenses, butoK > > rather than fix that Compaq chose to develop this program. This programlP > > doesn't even address that problem. It's clear from the stuff above that it'sM > > not permitted to use the base PAK from this program along with multi-user K > > PAKS from CSLG to achieve a multi-user system ( that would violate the i( > > "single, named individual" clause.)  > C > At least it now extends to all edu institutions, and specificallyrG > include students. But, I too feel that adding this and base licencingwC > to CSLG would be a much better move. Gateshit software is winninge@ > hand over fist in high schools here. And it's telling. A largeJ > local VMS/Alpha shop is planning to go to 'Newer Technologies' and Unix.3 > BTW, where _has_ Andrew been the last week or so?   G Personally, I am very concerned.  Very concerned.  Not for myself, but rD for Compaq (yes) and the VMS community at large.  Forget Andrew; it F would be nice to get some comments from the Q on this front.  If they D don't want to make any public statements, then surely someone could = email some trusted regular contributors to the newsgroup and C+ confidentially tell them where the beef is.c  G Several people have pointed out what needs to be done, and no-one from uI the Q has said why it isn't done.  We all know it CAN be done and SHOULD  " be done and probably MUST be done.  H Let's face it: the primary reason for the demise of VMS in some sectors D was that bad marketing took VMS out of academia, and thus people in I academia didn't get exposed to VMS and so a) there were fewer VMS people aH to hire (but at least it meant better salaries for those who were left) I and b) there were fewer people to buy VMS when they got in a position to jG make decisions and spend money.  There is now rough parity with regard 1D to hardware.  This COULD make for parity with regard to software if  implemented correctly.  G It looks like I'll be leaving academia soon, for financial reasons.  It>C looks like I'll get a job at a big VMS shop.  Lucky for me.  But ifoB things continue, shops will move away from VMS since there will beF no-one to hire, and people who don't know VMS won't push for it to be F bought.  PERSONALLY, I'm OK.  But I'm worried about the VMS community.   ------------------------------    Date: 08 Dec 2000 01:37:59 +0800/ From: Paul Repacholi <prep@k9.prep.synonet.com>g( Subject: Re: New OpenVMS  Education site0 Message-ID: <871yvkb094.fsf@k9.prep.synonet.com>  , helbig@astro.rug.nl (Phillip Helbig) writes:  J > Let's face it: the primary reason for the demise of VMS in some sectors F > was that bad marketing took VMS out of academia, and thus people in K > academia didn't get exposed to VMS and so a) there were fewer VMS people nJ > to hire (but at least it meant better salaries for those who were left) K > and b) there were fewer people to buy VMS when they got in a position to )I > make decisions and spend money.  There is now rough parity with regard hF > to hardware.  This COULD make for parity with regard to software if  > implemented correctly.   My spin on this is:)  ? Vaxes and VMS came out, but the edu discounts for VMS vanished.l@ **Everyone** wanted 780s, but the VMS hit was unexpected and too? much, so they searched for another OS. BSD was it. This boostedw> unix on the 11s already in place, and boosted BSD and the rest> of unix. Large numbers of Vax/BSD systems appeared from 80 on.  = Then DEC screwed over the 10/20s. Big time, and claimed Vaxenf; could do the job... Well, when the 8500 came along perhaps.s; But by then the momentum had built up. Graduates flowed outs; with experience at both app and system level with BSD/unix. < Other HW vendors redid the paging code to emulate a 780, and; bingo! an OS for the flood of RISC, Mini-Super and the restf= of the systems that poured out in the late 80s and early 90s.   > By then, teaching staff had only been in a unix ( and possibly> PC and Mac ) world, so what did they teach? And what did their5 students install when they got into the 'real world'?m  ; Now the VMS software needed is VERY thin on the ground, ando? another vendor has a huge chunk of the mindset in their pocket,l if not on the payroll.  = I've always thought that VMS took off in the comercial areanar@ because of the RSTS systems in schools and Unis. Where are their replacments now?  = The sad thing is, that in my experience, when people get somehB reasonable exposure to VMS, it a 'Wow! Why can't they do this?...'= Double that with Galaxy. And the common calling standard justi0 stuns the unix weenies. Not to mention clusters.  = It is a chicken and egg problem, and only the Q can break it.i9 And only by a LARGE jump, one that can't be hand waved or > talked down by the unix/gateshit hoard. And with a bottom line: that will be VERY painfull for the Uni managers to ignore.  @ The nice thing for the Q is that the cost can all be 'opertunity< cost'; they give up revinue that they may get. For a rapidly: vanishing value of 'may'! But, it will be a 5-10 year wait: for the pay back. IE, for the replacement to the Alpha and" what runs on that. VERY long term.  < The short term is that it will get SW vendors back on board.   ~Paul    ------------------------------  % Date: Thu, 07 Dec 2000 05:13:33 -0600r) From: Mike Drabicky <drabicky#dallas.net>N& Subject: Re: Pathworks to NT migration8 Message-ID: <fcqu2tgf8dk0tgun0368go0dk7ek1kkoe1@4ax.com>  # On Wed, 06 Dec 2000 16:40:21 -0300,i* fabio_compaq@ep-bc.petrobras.com.br wrote:  F >I am trying to migrate one  Pathworks file share to a WNT file share.= >I created a static name in the WINS of the PCSA file server,uE >example:  TEST, but its not mapping. I did it a few years ago but Is1 >dont remember if theres anything more (detail).t >u1 >When I try a NET VIEW \\TEST,  its not working.o > * >TEST is the alias of the new file server. > # >I know its a netbios question....o >t >Do  you have any idea ???  E Make sure that NetBEUI is running on your Pathworks server. You don'tn@ have to run it so it might not be active. That would explain the behavior you are seeing.  F Alternative: if you can see both your Pathworks and WinNT servers fromF a PC, just map the source and target shares on the PC. Once done, dragD and drop your source files over the destination files. Somewhat slowD (the PC sits in the middle of all this) but for a one time activity, not a big deal.o  
 Mike Drabicky    ------------------------------    Date: 07 Dec 2000 11:05:57 +0100G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>B5 Subject: Re: Problems installing MQ Series on VMS 6.2GH Message-ID: <y4itowh7ga.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>   kparris@my-deja.com writes:e  9 > > Any installation command file worth it's salt does a  I > > SET SYMBOL/SCOPE=NOGLOBAL as the second line. (So what is the first?)- > $ SET := SET   Bingo.   	Jan  < But then, I wouldn't have expected anything less from you...   ------------------------------  % Date: Thu, 07 Dec 2000 12:31:00 +0100   From: Paul Sture <paul@sture.ch>5 Subject: Re: Problems installing MQ Series on VMS 6.2 + Message-ID: <VA.000001ba.0df215a3@sture.ch>m  1 In article <90ms39$d7g$1@nnrp1.deja.com>,  wrote:n > From: kparris@my-deja.coms > Newsgroups: comp.os.vmse7 > Subject: Re: Problems installing MQ Series on VMS 6.2i% > Date: Thu, 07 Dec 2000 02:19:26 GMT8 > C > Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>b > wrote:< > > Any installation command file worth it's salt does a SET > SYMBOL/SCOPE=NOGLOBALg/ > > as the second line. (So what is the first?)l >  > $ SET := SET > # I think you got the cigar there :-)*  F This discussion sent me off to dig out Wolgang Moeller's SCHEDULE.COM  (that's the name I have it as).<  . He starts out with f$verify as the first line.  C Looking more closely at this procedure, when run in batch mode, he  J firstly sets the scope to global, and executes the commands passed in P2. E This will allow the batch job access to symbols defined in the login C command files.  G After executing the job commands, he sets scope to (nolocal,noglobal), $+ and this bit has me wanting to dig deeper. r   Here's that code snippet:t  
 $ set = "set" 
 $ set noon% $ set symbol/scope=(nolocal,noglobal)e $ if = "if"  $ delete = "delete"p $ delete/symbol set  $ delete/symbol if $ delete/symbol delete0 $ if f$type(goto).nes."" then delete/symbol goto0 $ if f$type(exit).nes."" then delete/symbol exit   &  (other commands used before the exit)  o  C Why is he doing this? The reason is that he executed the batch job 1H commands in line, so there may be local symbols set up within the scope  of _this_ procedure level.  H Global symbols may be set in those commands, but those will be excluded 6 by noglobal, so only local symbols need be considered.  J I am including Wolfgang's procedure here, as it's an excellent example of < the things to consider. It's also a quite useful utility :-)  J I will admit that I don't often go to these lengths, since I usually work D in a well known environment, but sometimes someone decides to add a 5 symbol to a group login file, and it can break stuff.e  
 Paul Sture Switzerlandt  D --------------------------------------------------------------------, Schedule.com as written by Wolfgang Moeller:D This will probably wrap, sorry about that - just look for lines not  starting with a dollar.s  + $ ver='f$verify(0,f$envir("verify_image"))'  $!9 $!*****    schedule commands for periodic batch executione1 $! author: w.j.m. jan 1990 (see at end of script)e2 $! modified 20-oct-1995 wjm: do not requeue after  JBC$_{JOBABORT,JOBREQUEUE}A $! modified 29-oct-1995 wjm: with /RESTART, do not requeue after o
 JBC$_QUERESET  $!5 $! INTERACTIVE (Startup) mode: prompt for everything,a; $! optionally: p1 = jobname,  p2..p5 = timespec /submitqual  $!/ $! BATCH mode: p1 = timespec, p2..p8 = commands G $! reuses all meaningful SUBMIT qualifiers except for /CLI, /PRIORITY,   /WSxxx:uD $! i.e. /CHAR /CPUTIME /noKEEP /noLOG /NAME /NOTIFY /noPRINT /QUEUE  /RESTART $! and all /PARAMETERS $!' $! NETWORK or OTHER mode: not supported  $! $! timespec: $!     AT daytime dayspecm $! or  EVERY deltatime $!6 $! dayspec: (comma-separated list, void implies daily)% $!     MON .. SUN  => on that weekdayd! $!     M1      => on 1st of months $! $!*****  $!
 $ set = "set"a% $ set symbol/scope=(nolocal,noglobal)n $ delete/symbol set2 $!+ $ if f$mode().nes."BATCH" then goto startupM $ delete/symbol ver  $!! $ call submittime_compute "''p1'"  $ gosub submittime_local $! $ sched_goto = "goto"s# $ on error then sched_goto resubmit  $! $ sched_np = 2 $loop:' $ if p'sched_np'.eqs."" then goto eloop4 $ sched_np = sched_np + 1n! $ if sched_np.lt.9 then goto loopj $eloop:I5 $ p'sched_np' = "sched_goto resubmit ! 'f$verify(0)'"  $!H $! live symbols are P1..P8, SCHED_NP, SCHED_SUBMITTIME, SCHED_GOTO (all  local) $! $ set symbol/scope=globalkC $ sched_np = sched_np + f$verify(1,0) * 0  ! fudge - use no commandp $ 'p2' $ 'p3' $ 'p4' $ 'p5' $ 'p6' $ 'p7' $ 'p8' $ 'p9', $ tmp = 'f$verify(0,0)'        ! "-"-catcher $!@ $resubmit: xstat = 'f$verify(0,0)'*0 + ($status .or. %x10000000) $!
 $ set = "set"U
 $ set noon% $ set symbol/scope=(nolocal,noglobal)e $ if = "if"a $ delete = "delete"e $ delete/symbol set- $ delete/symbol if $ delete/symbol delete0 $ if f$type(goto).nes."" then delete/symbol goto0 $ if f$type(exit).nes."" then delete/symbol exit0 $ if f$type(then).nes."" then delete/symbol then0 $ if f$type(else).nes."" then delete/symbol else2 $ if f$type(endif).nes."" then delete/symbol endif2 $ if f$type(gosub).nes."" then delete/symbol gosub4 $ if f$type(return).nes."" then delete/symbol return0 $ if f$type(show).nes."" then delete/symbol show2 $ if f$type(write).nes."" then delete/symbol write4 $ if f$type(submit).nes."" then delete/symbol submit $!2 $ p'sched_np' = ""         ! undo above assignment $!; $ JBC$_JOBABORT_INH =  %x10048084  ! (-F-) Qmgr deleting usrA $ JBC$_JOBREQUEUE_INH =    %x100480E4  ! (-F-) Qmgr requeueing usIC $ JBC$_QUERESET_INH =  %x100480EC  ! (-F-) Qmgr requeueing us with h /RESTART $!E $ if xstat.eq.JBC$_JOBABORT_INH .or. -     ! honor DELETE/ENTRY, etc.4F      xstat.eq.JBC$_JOBREQUEUE_INH then exit xstat  !!! get out of here $! $ restqual = "" 6 $ if f$getqui("display_job","job_restart",,"this_job") $ thenF $  if xstat.eq.JBC$_QUERESET_INH then exit xstat   !!! get out of here $  restqual = "/RESTART" $ endifl $!8 $ if f$getqui("display_job","job_cpu_limit",,"this_job") $ then $  cpulim = > f$edit(f$getqui("display_job","cpu_limit",,"this_job"),"trim"), $  if f$element(1," ",cpulim).nes." " then -H        cpulim = f$element(0," ",cpulim) + "-" + f$element(1," ",cpulim) ! $  cpuqual = "/CPUTIME=" + cpulimr $ else $  cpuqual = ""h $ endife $! $ notifyqual = ""r< $ if f$getqui("display_job","job_notify",,"this_job") then -    notifyqual = "/NOTIFY"v $!B $ charqual = f$getqui("display_job","characteristics",,"this_job")? $ if charqual.nes."" then charqual = "/CHAR=(" + charqual + ")"e $!< $ quename = f$getqui("display_job","queue_name",,"this_job") $!: $ jobname = f$getqui("display_job","job_name",,"this_job") $!7 $ if f$getqui("display_job","job_log_null",,"this_job")x $ then $  logqual = "/NOLOG"r $  printqual = ""g $ else $  logqual = "/LOG"tD $  logspec = f$getqui("display_job","log_specification",,"this_job"); $  if logspec.nes."" then logqual = logqual + "=" + logspeco $!9 $  if f$getqui("display_job","job_log_spool",,"this_job")- $  then- $      printqual = "/PRINT"0B $      printspec = f$getqui("display_job","log_queue",,"this_job")G $      if printspec.nes."" then printqual = printqual + "=" + printspectE $      if .not.f$getqui("display_job","job_log_delete",,"this_job") ->/            then printqual = printqual + "/KEEP"a $  elseM $      printqual = "/NOPRINT"i@ $      if f$getqui("display_job","job_log_delete",,"this_job") -1            then printqual = printqual + "/NOKEEP"s $  endif $ endif  $!* $ write SYS$OUTPUT f$fao("!/----------!/") $! $ gosub qquote_submitpar( $ submit/queue='quename'/name=&jobname -G    'restqual' 'cpuqual' 'notifyqual' 'logqual' 'printqual' 'charqual' -o     /after=&sched_submittime    -0    /param=(&p1,&p2,&p3,&p4,&p5,&p6,&p7,&p8)    -&    'f$parse(";",f$envir("procedure"))' $ exit $status $! $!@ $qquote_submitpar: ! GOSUB - double `"'s in P1 .. P8 and JOBNAME $ p9 = jobname !fudget $ qq_np = 0> $qq_np_loop: $ qq_np = qq_np + 1d$ $ if qq_np.gt.9 then goto qquote_end $ if p'qq_np'.eqs."" $ then9 $  p'qq_np' = """"""   ! for &<pn>, <pn> must be non-voidw1 $  goto qq_np_loop     ! (but may be blanks etc.)e $ endift8 $ if p'qq_np' - """" .eqs. p'qq_np' then goto qq_np_loop
 $ qq_i = 0
 $ qq_tmp = ""n	 $qq_loop:e& $ qq_tmp1 = f$elem(qq_i,"""",p'qq_np')+ $ if qq_tmp1.eqs."""" then goto qq_loop_endo) $ qq_tmp = qq_tmp + """" + qq_tmp1 + """"y $ qq_i = qq_i + 1  $ goto qq_loop
 $qq_loop_end:  $ p'qq_np' = qq_tmpo $ goto qq_np_loopt $qquote_end: $ jobname = p9 !fudge  $ delete/symbol p9 $ return $! $!D $submittime_local: ! GOSUB, make global SCHED_SUBMITTIME into local  symbol
 $ set = "set"  $ delete = "delete"n$ $ sl_scope = f$envir("symbol_scope")% $ set symbol/scope=(nolocal,noglobal)tH $ if f$type(sched_submittime).nes."" then delete/symbol sched_submittime $ set symbol/scope=global+% $ sched_submittime = sched_submittimez' $ delete/symbol/global sched_submittime  $ set symbol/scope=('sl_scope')o $ delete/symbol sl_scope $ delete/symbol setg $ delete/symbol delete
 $ return 1 $! $!= $submittime_compute: subroutine    ! p1 = timespec, output = m SCHED_SUBMITTIME
 $ set = "set"l% $ set symbol/scope=(nolocal,noglobal)o. $ on warning then exit $status .or. %x10000000 $!" $ right_now = f$cvtim(,"absolute"). $ today = f$cvtim(right_now,"absolute","date") $!I $ SHR$_BADTIME = %x00001194    ! "'!AS' is an invalid time specification"s $!3 $ daypattern = "\\MON\TUE\WED\THU\FRI\SAT\SUN\M1\\"  $!. $ timespec = f$edit(p1,"trim,upcase,compress")% $ timekey = f$element(0," ",timespec)t? $ if f$element(1," ",timespec).eqs." " then return SHR$_BADTIMEt' $ if timekey.eqs."AT" then goto time_atn- $ if timekey.eqs."EVERY" then goto time_deltat $ return SHR$_BADTIMEb $! $time_delta:5 $ deltatime = f$cvtim(timespec-(timekey+" "),"delta") ? $ deltatime = f$element(0," ",deltatime) + "-" + f$element(1," ' ",deltatime)@ $ sched_submittime = f$cvtim(right_now+"+"+deltatime,"absolute") $ goto time_ok $!	 $time_at:i: $ daytimespec = f$cvtim(f$element(1," ",timespec),"delta")B $ if f$element(0," ",daytimespec).nes."0" then return SHR$_BADTIME' $ hhmmss = f$element(1," ",daytimespec)i% $ dayspec = f$element(2," ",timespec)d) $ if dayspec.eqs." " then goto time_dailys $! parse dayspec
 $ daymask = 0o $ i = 0m $daykeyloop:# $ daykey = f$element(i,",",dayspec)a+ $ if daykey.eqs."," then goto daykeyloopende) $ j = f$locate("\"+daykey+"\",daypattern)lC $ if j.eq.f$length(daypattern) .or. (j.and.(4-1)).ne.1 then return d SHR$_BADTIME $ daymask[j/4,1] = 1 $ i = i + 1m $ goto daykeyloope $daykeyloopend:t@ $ if daymask.eq.0 then goto time_daily         ! must not happen2 $ if (daymask.and.127).eq.0 then goto time_monthly $! $ weekdaymask = 0i $ weekday = today < $ if f$cvtim(right_now).ges.f$cvtim(today+":"+hhmmss) then -4    weekday = f$cvtim(today+"+1-0","absolute","date")
 $weekdayloop:CF $ daykey = f$edit(f$extract(0,3,f$cvtim(weekday,,"weekday")),"upcase")) $ j = f$locate("\"+daykey+"\",daypattern)  $ weekdaymask[j/4,1] = 1< $ if (weekdaymask.and.daymask).ne.0 then goto weekdayloopend5 $ weekday = f$cvtim(weekday+"+1-0","absolute","date")i $ goto weekdayloop $weekdayloopend:9 $ if (daymask.and.(.not.127)).eq.0 then goto time_weekdayi $ gosub monthday_computeC $ if f$cvtim(monthday).lts.f$cvtim(weekday) then goto time_monthdaye $time_weekday:+ $ sched_submittime = weekday + ":" + hhmmssw $ goto time_ok $! $monthday_compute: ! GOSUBE $! need not test "daymask" since there is currently only 1 case: "M1"nG $ j = 32 - f$integer(f$cvtim(today,,"day"))    ! #days to get to "32nd"kB $ monthday = f$cvtim(today+"+"+f$string(j)+"-0","absolute","date")7 $ monthday = "1" + (monthday-f$element(0,"-",monthday))l $ return       ! from GOSUBl $! $time_monthly: $ gosub monthday_compute $time_monthday:c, $ sched_submittime = monthday + ":" + hhmmss $ goto time_ok $! $time_daily:E $ sched_submittime = f$cvtim(today+"+1-0","absolute","date") + ":" + , hhmmss $ goto time_ok $!	 $time_ok:  $ set symbol/scope=global-& $ sched_submittime == sched_submittime $ set symbol/scope=noglobalc $ exit 1 $ endsubroutinec $! $! $!*****    non-batch startup $!	 $startup:g# $ on warning then goto startup_exit.? $ if f$mode().nes."INTERACTIVE" then return %x354  ! SS$_IVMODEa $! $ write SYS$OUTPUT -4 "Schedule commands for periodic BATCH execution ..." $! $ jobname = p1* $ if jobname.nes."" then goto inq_name_end
 $inq_name: $ read SYS$COMMAND jobname - /prompt="Job name: "& $ if jobname.eqs."" then goto inq_name $inq_name_end: $!2 $ cmdline = f$edit(p2+" "+p3+" "+p4+" "+p5,"trim") $ cmdqual = ""& $ if cmdline.eqs."" then goto inq_mode% $ timespec = f$element(0,"/",cmdline)o+ $ cmdqual = f$edit(cmdline-timespec,"trim") ' $ p1 = f$edit(timespec,"compress,trim") ! $ if p1.eqs."" then goto inq_model! $ call submittime_compute "''p1'"  $ gosub submittime_local $ starttime = sched_submittime $ goto inq_cmd $!
 $inq_mode: $ read SYS$COMMAND answer -d. /prompt="Schedule by INTERVAL or by DAYTIME: "' $ answer = f$edit(answer,"upcase,trim")l6 $ if f$extract(0,1,answer).eqs."I" then goto inq_delta4 $ if f$extract(0,1,answer).eqs."D" then goto inq_day $ goto inq_moded $! $inq_delta:  $ read SYS$COMMAND deltatime -( /prompt="Time interval ([dd-]hh[:mm]): "( $ p1 = "EVERY "+f$edit(deltatime,"trim")  $ on warning then goto inq_delta! $ call submittime_compute "''p1'" # $ on warning then goto startup_exitW $ gosub submittime_local $ goto inq_starttime $!	 $inq_day:  $ read SYS$COMMAND daytime -! /prompt="Time of day (hh[:mm]): "s, $ p1 = "AT "+f$edit(daytime,"collapse,trim") $ on warning then goto inq_day! $ call submittime_compute "''p1'"y# $ on warning then goto startup_exity $ gosub submittime_local $ answer = 0 $ read SYS$COMMAND answer -rH /prompt="Do you want to restrict scheduling to certain days? (Y or N): "" $ if answer then goto inq_dayspec1 $ goto inq_starttime $! $inq_dayspec1:
 $ p1s = p1 $ write SYS$COMMAND -h= "You may select particular days by weekday, or 1st of month."T $ write SYS$COMMAND -d2 "Choices are: MON,TUE,WED,THU,FRI,SAT,SUN and M1."
 $inq_dayspec:i $ read SYS$COMMAND dayspec -, /prompt="Selection (comma-separated list): "+ $ dayspec = f$edit(dayspec,"trim,collapse")i $ if dayspec.eqs.""m $ then6 $  write SYS$COMMAND "No selection - DAILY scheduling" $  goto inq_starttime  $ endifn $ p1 = p1s + " " + dayspec" $ on warning then goto inq_dayspec! $ call submittime_compute "''p1'"d# $ on warning then goto startup_exitt $ gosub submittime_local $ goto inq_starttime $!7 $inq_starttime:        ! so far, we have JOBNAME and P1f $ read SYS$COMMAND starttime -; /prompt="First start time (default: ''sched_submittime'): "e7 $ if starttime.eqs."" then starttime = sched_submittimeb $!	 $inq_cmd:a	 $ p2 = "" 	 $ p3 = ""E	 $ p4 = ""e	 $ p5 = ""o	 $ p6 = ""a	 $ p7 = ""n	 $ p8 = ""V $ write SYS$COMMAND -h0 "Enter commands to be executed (RETURN if done)" $ np = 2 $inq_cmd_loop: $ p'np' = "" $ read SYS$COMMAND p'np' -% /prompt="Command ''f$string(np-1)': "i $ p'np' = f$edit(p'np',"trim")+ $ if p'np'.eqs."" then goto inq_cmd_loopendD
 $ np = np + 1t# $ if np.le.8 then goto inq_cmd_loopW $inq_cmd_loopend:4 $!# $ defaultqual = "/RESTART /NOPRINT"s8 $ if f$parse(jobname,"SYS$LOGIN:.LOG;").eqs."" then    -(    defaultqual = defaultqual + " /NOLOG" $ submitqual = cmdqual/ $ if submitqual.nes."" then goto startup_submitt $ write SYS$COMMAND -d1 "Default SUBMIT qualifiers will be: ",defaultqual $! $ write SYS$COMMAND - E "Enter (optional) SUBMIT qualifiers (e.g. /QUEUE,/CPU,/PRINT,/noLOG)"e $ read SYS$COMMAND submitqual -  /prompt="Qualifier(s): "( $ submitqual = f$edit(submitqual,"trim")H $ if submitqual.nes."" .and. f$extract(0,1,submitqual).nes."/" then    -     submitqual = "/" + submitqual $! $startup_submit: $ gosub qquote_submitpar( $ submit 'defaultqual' /name=&jobname  -    /after=&starttime   -    'submitqual'    -0    /param=(&p1,&p2,&p3,&p4,&p5,&p6,&p7,&p8)    -'    'f$parse(";",f$envir("procedure"))' f $! $startup_exit:I $ exit $status .or. %x10000000 + f$verify(ver,f$envir("verify_image")) *   0n $i5 $!Wolfgang J. Moeller, Tel. +49 551 2011516 or -510, - moeller@gwdvms.dnet.gwdg.de G $!GWDG, D-37077 Goettingen, F.R.Germany   |       Disclaimer: No claim r	 intended!.< $!<moeller@decus.decus.de>  ----- <moeller@gwdg.de>  -----   <w.moeller@ieee.org>   ------------------------------  % Date: Thu, 07 Dec 2000 15:42:05 -0200i1 From: "Valdemir J. Santos" <valdemir-@uol.com.br>a# Subject: RAID array informations... 5 Message-ID: <000f01c06075$05f423a0$c51ebfc8@valdemir>i  F I=B4d like to get more informations about RAID arrays and yours compl= exity.  E Where I can get RAID arrays dummie=B4s informations in the internet ?   F Where I can get the book "The RAIDBOOK - A source for RAID Technology= " ?:   Thank you in advance.e   ------------------------------  % Date: Wed, 06 Dec 2000 08:22:47 -0700e+ From: "Barry Treahy, Jr." <Treahy@mmaz.com>r Subject: Re: raid system neededr( Message-ID: <3A2E59C7.A7879B81@mmaz.com>  $ hoffman@xdelta.zko.dec.nospam wrote:  k > In article <000c01c05ef7$46a89400$2c96a8c6@mscmain.com>, Hank Vander Waal <hvanderw@novagate.com> writes: H > :I have a customer with 1000a 5/400 that needs to add an external raidM > :system.  He has about 15 drives that are in the SBB drive  containers that M > :he would like to use.  Can someone tell me what raid box I can get for him  > :??  (running VMS 7.1)  l Check out Infortrends RAID controllers, they have pretty slick SCSI to SCSI RAID controllers.  I used one ofl their VAR's, Media Intergration out in California, but regardless, their controllers have four channels thatg can be configured as host or drive and the storage see on the drive channels can be sliced and diced inoC various RAID configurations.  Very flexible, fast and inexepensive!o   Barryl     --  ? Barry Treahy, Jr  *  Midwest Microwave  *  Vice President & CIOs  A E-mail: Treahy@mmaz.com * Phone: 480/314-1320 * FAX: 480/661-7028o   ------------------------------  " Date: Thu, 7 Dec 2000 10:05:38 GMT( From: Terry Kennedy <terry@gate.tmk.com>2 Subject: Re: Source/availability of 60m DAT media?' Message-ID: <G5701E.MnJ@spcuna.spc.edu>x  " l_ricker@lto.locktrack.com writes:G > Does anyone know of a last-gasp media supplier, or have a dozen or sob? > extras to spare at reasonable cost?  Responses to this can gop> > offline/private so's not to clog up the main threads... TIA!  F   I never had any 60M-only drives, but I get all my DAT media from TheB Tape Company (http://www.thetapecompany.com/). You could try them.  4         Terry Kennedy             http://www.tmk.com5         terry@tmk.com             Jersey City, NJ USA)   ------------------------------  $ Date: Thu, 7 Dec 2000 12:27:30 -0500) From: "Ebinger . Eric" <EEbinger@drc.com>t2 Subject: RE: Source/availability of 60m DAT media?B Message-ID: <7162F87E9EF4D311BA9900805FC1D3AE7A625A@and02.drc.com>  3 CDW (Computer Discount Warehouse) lists 60M 4mm DAT- tapes...  7 http://www.cdw.com/shop/products/default.asp?EDC=164167r     Eric Ebinger     > -----Original Message-----F > From: l_ricker@lto.locktrack.com [mailto:l_ricker@lto.locktrack.com], > Sent: Wednesday, December 06, 2000 2:02 PM > To: Info-VAX@Mvb.Saic.Coma0 > Subject: Source/availability of 60m DAT media? >  > : > I've got a couple of old TLZ06 (or older?) 4mm DAT tape  > drives to use for a9@ > few more months (or years?...), but am running out of media.   > These drives> > are not on "mission-critical" systems, and not on a network  > (for device-7 > sharing), but it's necessary to continue to use them.e > ? > Efforts to order 60m (that's right, sixty-meter, aka "DDS1") e > media from? > Compaq and elsewhere have made my purchasing guy crazy... he N > keeps hearingI= > "obsolete, no longer made".  Nope, these drives will *not*   > accept 90m oruH > longer media, thanks to built-in pre-obsolescence features like "Media > Recognition System". > G > Does anyone know of a last-gasp media supplier, or have a dozen or soa? > extras to spare at reasonable cost?  Responses to this can go > > offline/private so's not to clog up the main threads... TIA! > : > Lorin Ricker                    Lorin.Ricker@t-netix.com > T-NETIX, Inc. 8 > 67 Inverness Dr. East           (303)705-5575 (direct) > Englewood, CO 80112" >    ------------------------------  % Date: Thu, 07 Dec 2000 18:59:38 +0100p( From: Etienne Vogt <vogt@alice.obspm.fr>2 Subject: Re: Source/availability of 60m DAT media?. Message-ID: <3A2FD00A.B957916A@alice.obspm.fr>  ! l_ricker@lto.locktrack.com wrote:yI > Efforts to order 60m (that's right, sixty-meter, aka "DDS1") media from L > Compaq and elsewhere have made my purchasing guy crazy... he keeps hearingJ > "obsolete, no longer made".  Nope, these drives will *not* accept 90m orH > longer media, thanks to built-in pre-obsolescence features like "Media > Recognition System".  G 90m tapes are also DDS1, they have the same recording format as the 60me7 ones, DDS2 uses 120m tapes and DDS3 124m tapes I guess. G It seems strange to me that your drive would not accept 90m tapes, as ItH have used some pretty old DDS1 drives and they have all accepted to work
 with them.F On the other hand, drives with "Media Recognition System" enabled willF refuse to write on audio DAT tapes or on early 60m DDS tapes that lackC the MRS stripes on the tape lead, but they may have a DIP switch tof disable this behaviour.e   -- e# 		Etienne Vogt (vogt @ na.astro.it).) 		Osservatorio Astronomico di Capodimontet 		Napoli, Italia   ------------------------------  % Date: Thu, 07 Dec 2000 08:30:39 -0700   From: l_ricker@lto.locktrack.com. Subject: Source/availability of 60m DAT media?. Message-ID: <00120708303967@lto.locktrack.com>  L Thanks to all who've responded!  Terry, your reference to thetapecompany.comE is very much appreciated... I've put my purchasing guy on that trail!e  L Turns out: `my bad'... I'd neglected to try (for lack of having one in-hand)L a 90m DAT in my old drive, and that turns out to work just as well for me asJ it does for the rest of you, and since 90's are still more-or-less readily5 available, my "crisis" has past.  Again, many thanks.    Lorina   ------------------------------  % Date: Thu, 07 Dec 2000 05:02:00 -0800", From: Jack Patteeuw <jjpatteeuw@voyager.net> Subject: Re: Sun Cluster8 Message-ID: <3a2f5fac$0$30004$2c3edae7@news.voyager.net>   Bill Todd wrote:  uE > Compaq now has well over two years' worth of its own negligible VMSuK > promotion to answer for, and unlike BP are still in a position to be helduF > accountable for the disparity between their encouragement of the VMS> > faithful and near-total neglect of VMS in any wider context.   <snip>  A Hmmm ... Interesting thought went through my mind.  Why not sue ?=  > Major headlines in the Detroit area about Kirk Kerkorian suingE DaimlerChrysler after CEO Juergen Schrempp admitted publicly that he ,C had lied about the takeover being a "merger of equals".  The claim oE in the lawsuit is that Schrempp's statement had devalued a corporate g> asset (Chrysler division) causing stock holders to lose money.  ; Any Compaq stock owners want to start a class action suit ?aB (You could become a stock owner real cheap these days !!  Although2 one stock analyst had Compaq as a "buy" recently")    
 Jack Patteeuwc   ------------------------------  % Date: Thu, 07 Dec 2000 05:15:57 -0800., From: Jack Patteeuw <jjpatteeuw@voyager.net> Subject: Re: Sun Cluster8 Message-ID: <3a2f62f1$0$30009$2c3edae7@news.voyager.net>   jbecker@ui.urban.org wrote:  > I > At the November 28 ESILUG meeting (Washington DC area), we heard a good_F > talk comparing "cluster" features on various technologies offered by > Compaq and others.   <snip>  G Interesting data.  One thing I did not see mentioned in the spreadsheetbB that accompanied the presentation was any references to "locking".  B My personal opinion is that the VMS Distributed Lock Manager sets > VMSclusters significantly apart from any other "cluster".  DLME and ISAM files gives "most" of the functionality of a rdb (admittedlyqD with only one table) at zero cost.  This is a strategic advantage !!  
 Jack Patteeuw    ------------------------------   Date: 7 Dec 2000 11:29:31 +0100n* From: eplan@kapsch.net (Peter LANGSTOEGER) Subject: Re: Sun Cluster* Message-ID: <3a2f668b$1@news.kapsch.co.at>  E In article <90mtnk$eb5$1@nnrp1.deja.com>, kparris@my-deja.com writes:"C >From where I sit, Compaq has been significantly more supportive ofqD >(and less destructive to) VMS than Digital was in the latter years.   Indeed. But is this enough ?  E >You could almost see the light-bulbs go on when the top-level Compaq I >management realized how valuable the profitable VMS business was to a PCoE >company with razor-thin margins, and the puzzlement as to why in theoI >world Digital had been actively trying to kill it off.  I've seen thingst8 >to be very different from that inflection point onward.  4 Then why do they still destroy the brand "Digital" ?  C >In the installed base in general, we've gotten beyond the phase ofyH >everyone feeling forced to move quickly off of VMS because they felt itI >would be gone in 5 years.  Today, most existing customers feel confidentoG >that even if they do nothing, they are safe in staying with VMS.  Some I >attrition is still occurring, but it is being met and even exceeded withUI >new accessions.  I'd like to get quickly beyond that to a faster-growingi* >customer base, but we've come a long way.  = And this I meant. A long way in the wrong direction was gone.sI Now all must be forced to run as fast as they can to the right direction,i/ or the loose the contact to the "leading edge".\E And I see them only walking (and some now doesn't see them at all)...p   --  < Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111-2651; Network and OpenVMS system manager  Fax.    +43 1 81111-888"< <<< KAPSCH AG  Wagenseilgasse 1     E-mail  eplan@kapsch.netH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"   ------------------------------   Date: 7 Dec 2000 08:33:55 -0500t9 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)i Subject: Re: Sun Cluster+ Message-ID: <OaZTfyh8BpC6@eisner.decus.org>t  g In article <3a2f62f1$0$30009$2c3edae7@news.voyager.net>, Jack Patteeuw <jjpatteeuw@voyager.net> writes:. >  >  > jbecker@ui.urban.org wrote:= >> eJ >> At the November 28 ESILUG meeting (Washington DC area), we heard a goodG >> talk comparing "cluster" features on various technologies offered by  >> Compaq and others.k >  > <snip> > I > Interesting data.  One thing I did not see mentioned in the spreadsheet"D > that accompanied the presentation was any references to "locking". > D > My personal opinion is that the VMS Distributed Lock Manager sets @ > VMSclusters significantly apart from any other "cluster".  DLMG > and ISAM files gives "most" of the functionality of a rdb (admittedlynF > with only one table) at zero cost.  This is a strategic advantage !!  H I think the real advantage is that VMS programming discipline has causedG most programs (thanks to compiler/RTL construction) to actually respectcE the Lock Manager, so programs in a cluster cooperate "automatically".+  N ==============================================================================N Great Inventors of our time: Al Gore -> Internet; Sun Microsystems -> ClustersN ==============================================================================   ------------------------------  # Date: Thu, 07 Dec 2000 16:06:26 GMTc From: jbecker@ui.urban.org Subject: Re: Sun Cluster) Message-ID: <90ochs$if3$1@nnrp1.deja.com>s  8 In article <3a2f62f1$0$30009$2c3edae7@news.voyager.net>,/   Jack Patteeuw <jjpatteeuw@voyager.net> wrote:1  = > Interesting data.  One thing I did not see mentioned in thea spreadsheet"D > that accompanied the presentation was any references to "locking".  B He discussed it in the presentation with some emphasis. There is aD column in the spreadsheet headed "Cluster File System" which impliesE the presence or absence of DLM, but it does seem like an extra columnt% that addresses locking would be nice.t   --
 Jim Becker+ The Urban Institute (http://www.urban.org/)s7 Encompass ESILUG (http://eisner.decus.org/lugs/esilug/)a    & Sent via Deja.com http://www.deja.com/ Before you buy.e   ------------------------------  % Date: Thu, 07 Dec 2000 15:27:10 +0000p% From: Alan Greig <agreig@my-deja.com>o Subject: Re: Sun Cluster8 Message-ID: <tmav2tc0n9dpjojcs3vp6sgdeorbn74j45@4ax.com>  B On 7 Dec 2000 11:29:31 +0100, eplan@kapsch.net (Peter LANGSTOEGER) wrote:    D >>In the installed base in general, we've gotten beyond the phase ofI >>everyone feeling forced to move quickly off of VMS because they felt ithJ >>would be gone in 5 years.  Today, most existing customers feel confidentH >>that even if they do nothing, they are safe in staying with VMS.  SomeJ >>attrition is still occurring, but it is being met and even exceeded withJ >>new accessions.  I'd like to get quickly beyond that to a faster-growing+ >>customer base, but we've come a long way.  >u> >And this I meant. A long way in the wrong direction was gone.J >Now all must be forced to run as fast as they can to the right direction,0 >or the loose the contact to the "leading edge".F >And I see them only walking (and some now doesn't see them at all)...  B It seems to me as if Rich Marcello has been given the authority toF push VMS but other parts of Compaq don't seem to have got the message.@ The question is is this intentional on Compaq's part, due to badF internal communication or due to internal unauthorized 'enemy action'?B Whatever the reason it is currently seriously harming all the good% work put in by Marcello and his team.C --
 Alan Greig   ------------------------------  % Date: Thu, 07 Dec 2000 15:31:31 +0000,% From: Alan Greig <agreig@my-deja.com>  Subject: Re: Sun Cluster8 Message-ID: <i6bv2tcf4sbncpch6k1vkcdivtf5i36fk8@4ax.com>  F On 7 Dec 2000 08:33:55 -0500, Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) wrote:s    O >==============================================================================qO >Great Inventors of our time: Al Gore -> Internet; Sun Microsystems -> ClusterscO >==============================================================================s  = I think you'll find it was actually the Al Gore Rythm your VPd	 invented.D  ? Yes I know it's been posted elsewhere but I still find it funny    --
 Alan Greig   ------------------------------  % Date: Thu, 07 Dec 2000 16:10:36 +0000l  From: steven.reece@quintiles.com Subject: Re: Sun ClusterH Message-ID: <OF9137BA54.6D705DA7-ON802569AE.00588DB4@qedi.quintiles.com>  2 Peter Langstoeger (eplan at kapsch dot net) asked:C >>>Then why do they (Compaq) still destroy the brand "Digital" ?<<<s  F Because it's not Digital any more.  It is Compaq.  One big Compaq thatI offers the whole range of products from PCs through Wintel servers, Linuxf boxes, Tru64, NSK and OpenVMS.K Compaq are not a PC box shifter, they are a company offering the full range5. of product for the full range of applications.   Steve.   ------------------------------  % Date: Thu, 07 Dec 2000 14:57:04 -0300c) From: fabio_compaq@ep-bc.petrobras.com.brp Subject: Re: Sun ClusterL Message-ID: <OFD322A606.13A9641A-ON032569AE.00626CD1@ep-bc.petrobras.com.br>  E Sometimes I feel, OpenVMS should be an independent group from Compaq.WK Like the linux.org....to develop and promote the OS. If Compaq permit otherj& companies develop Alpha hardware......       Regardsu   FC              6 Alan Greig <agreig@my-deja.com> em 07/12/2000 13:27:10L                                                                             L                                                                             L                                                                                 >                                                               >                                                               >                                                               >  Para:    Info-VAX@Mvb.Saic.Com                               >                                                               >  cc:      (bcc: Fabio dos Santos Cardoso/E-P-BC/Contratada)   >                                                               >                                                               >                                                               >  Assunto: Re: Sun Cluster                                     >                                                                             B On 7 Dec 2000 11:29:31 +0100, eplan@kapsch.net (Peter LANGSTOEGER) wrote:    D >>In the installed base in general, we've gotten beyond the phase ofI >>everyone feeling forced to move quickly off of VMS because they felt it J >>would be gone in 5 years.  Today, most existing customers feel confidentH >>that even if they do nothing, they are safe in staying with VMS.  SomeJ >>attrition is still occurring, but it is being met and even exceeded withJ >>new accessions.  I'd like to get quickly beyond that to a faster-growing+ >>customer base, but we've come a long way.  >,> >And this I meant. A long way in the wrong direction was gone.J >Now all must be forced to run as fast as they can to the right direction,0 >or the loose the contact to the "leading edge".F >And I see them only walking (and some now doesn't see them at all)...  B It seems to me as if Rich Marcello has been given the authority toF push VMS but other parts of Compaq don't seem to have got the message.@ The question is is this intentional on Compaq's part, due to badF internal communication or due to internal unauthorized 'enemy action'?B Whatever the reason it is currently seriously harming all the good% work put in by Marcello and his team.n --
 Alan Greig   ------------------------------  % Date: Thu, 07 Dec 2000 08:12:51 -0800,+ From: "richard n. frank" <rnfrank@llnl.gov>a2 Subject: Sun CSV Constitution Election Belly Laugh> Message-ID: <4.2.0.58.20001207080047.00b06db0@poptop.llnl.gov>  
 Good morning, D I thought I'd run this up the flagpole and see if anybody salutes...  F Has anyone ever considered restructuring this group along the model of5 tru64-unix-managers@ornl.gov?  That is a great group.-% People submit questions to the group,h- responses are directed to the submitter only, A the submitter is obliged to summarize responses back to the list.e  H There is a lot of good information in this group too, but the volume of  mail is staggering.n   Thanks for the thoughts.*                                       rich  @ DISCLAIMER: This is not the opinion of the U.S. Government, DOE,2 UC, or LLNL. It may not even really be my opinion.   ------------------------------  % Date: Thu, 07 Dec 2000 17:52:18 +0000a- From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>u6 Subject: Re: Sun CSV Constitution Election Belly Laugh) Message-ID: <3A2FCE51.3B12A63C@bbc.co.uk>    "richard n. frank" wrote:    > Good morning,-F > I thought I'd run this up the flagpole and see if anybody salutes... >"H > Has anyone ever considered restructuring this group along the model of7 > tru64-unix-managers@ornl.gov?  That is a great group..' > People submit questions to the group,u/ > responses are directed to the submitter only,wC > the submitter is obliged to summarize responses back to the list.- >OI > There is a lot of good information in this group too, but the volume oft > mail is staggering.t  D I have used the alpha-osf-managers  in the past, and liked it, for a! unix problem solving environment.    however, points againstg  F (1) a lot of people read this group viua USENET not email. Suggest you? try that if your mailbox is overflowing with INFO-VAX postings.   = (2) Any change of newsgroup name may not be propogated to alltA organizations. eg I can still get comp.unix.osf.osf1 here but noto comp.unix.tru64,J and the news service managfer has no interest in helpinmg me out with this (IE am not working with Tru64 but would be interested in browsing once inN	 a while).e  J (3) Sometimes threads get divergent but then that can be interesting. OTOH@ US elections and AH threads have been somewhat tedious recently.  E Tim, who works very hard at refraining from TOO much personal opiniona) on the net these days, life is too short.s   --6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uki  A I speak for myself only and my views in no way represent those of  MedAS or the BBC.c   ------------------------------    Date: 07 Dec 2000 11:20:58 +0100G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>mR Subject: Re: Sun new cluster comparison ommitted VMS (but VMS was their benchmark)H Message-ID: <y4g0k0h6r9.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  , kuhrt@eisner.decus.org (Marty Kuhrt) writes:  = > On another note, how does the patent for Sun's "Global filecC > system-based system and method for rendering devices on a clusterdF > globally visible", US6047332, differ from how VMS has been doing it?  N I don't think VMS uses a single registrar to manager the device name database.N Dunno about the new naming scheme for SCSI clusters...I have the impression itL is in the system manager's hands to keep the descriptions on different nodes consistent.a   	Jan   ------------------------------   Date: 7 Dec 2000 08:38:51 -0500t9 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)eR Subject: Re: Sun new cluster comparison ommitted VMS (but VMS was their benchmark)+ Message-ID: <5d+IuuoaDf59@eisner.decus.org>p   In article <y4g0k0h6r9.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>, Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de> writes:. > kuhrt@eisner.decus.org (Marty Kuhrt) writes: > > >> On another note, how does the patent for Sun's "Global fileD >> system-based system and method for rendering devices on a clusterG >> globally visible", US6047332, differ from how VMS has been doing it?s > P > I don't think VMS uses a single registrar to manager the device name database.  A By default, VMS uses algorithmic naming, so there is no conflict.n  P > Dunno about the new naming scheme for SCSI clusters...I have the impression itN > is in the system manager's hands to keep the descriptions on different nodes
 > consistent.l  8 Yes, SCSI clusters and DSSI clusters require management.  N ==============================================================================N Great Inventors of our time: Al Gore -> Internet; Sun Microsystems -> ClustersN ==============================================================================   ------------------------------    Date: 08 Dec 2000 00:57:56 +0800/ From: Paul Repacholi <prep@k9.prep.synonet.com>d( Subject: Testers wanted for Spell update0 Message-ID: <874s0gb23v.fsf@k9.prep.synonet.com>  0 Hi folks, I am looking for a few people to trial1 an update to Spell before it goes to the Freeware  CD. Alpha and Vax, 7.2.g   ~paulh   ------------------------------  # Date: Thu, 07 Dec 2000 11:02:09 GMTu+ From: "Andy Proctor" <aproctor@hotmail.com> D Subject: Trying to get Sevice and Support for a Couple of VMS Alphas4 Message-ID: <976186928.554972@ananke.eclipse.net.uk>   All,  G Can someone from COMPAQ reading this newsgroup please help. I have been H trying to purchase a hardware support agreement for one of my customers,L located in Holland. COMPAQ in Holland seem totally unable to provide me withL a even a quotation for the equipment. Apparemtly they have given the detailsC to a thrid party organisation, who have not to date contacted me as0H requested. I have been attempting to get a quotation for nearly 2 monthsL now, this really is not the way a company should behave do you not think? No. wonder people are moving from COMPAQ products.K Would anyone from COMPAQ be able to to use some influence, or indeed put mecI in touch with a service and support organisation in Holland, (I am in thedG UK) so that I can obtain a quote and then purchase a support agreement!d  & Thanks in advance to all you may help. Frustrated Regards   Andy Proctor   ------------------------------  $ Date: Thu, 7 Dec 2000 16:24:26 +0100% From: "Philip J. Lewis" <phl@bwsc.dk>  Subject: Re: US Constitution0 Message-ID: <oYNX5.170$bK3.3874@news.get2net.dk>  ! John Macallister wrote in message C <35666012DF4CD411BE940090279FA240112000@ppnt41.physics.ox.ac.uk>...uL >>Well, given the absolute numbers of soldiers deployed on the eastern frontF >>by Germany (all having their balls frozen off), v. the number in theJ >>European theatre, I think we should give the "saved our bacon" credit to >theL >>Russians (yes I know about Murmansk).  The Americans, and all the rest did. >>help though, and quite important it was too! >SL >This in itself is a distortion of the WWII story. If it hadn't been for theJ >NAZI-Russian alliance at the beginning Hitler may not have invaded PolandK >when he did and this alliance freed up soldiers to start a campaign in thew >West. >h  * I do not see the connection - call me dim.  L >While absolute numbers are often impressive the impact on the final outcomeL >may not always be proportional to the numbers involved. The real saviour ofF >this country was the English Channel just as the Pacific and Atlantic OceansG >and their vastness  ensured the USA's continued prosperity. It was theo; >vastness of "Russia" which made the real difference there., >s  G Indeed - Mother Russia and her size (and brutal weather).  War is aboutrC geography, and this, along with the channel are very good examples.tL However, the resources expended in the East did have a serious effect of the> ability of the Germans to effectively deal with everyone else.  F >It would have taken just one bullet in WWII to end the whole businessI >however many were used overall. It could take just one vote to decide andH >election however many votes were cast and accurately counted. As we all knowJ >it could take just one bit to make the difference between a program which >works and one which doesn't.4   Maybe - we will never know.n   >y >Johnn >rC >Name: John B. Macallister  E-mail: j.macallister1@physics.ox.ac.uklI >Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UKtB >Phone: +44-1865-273388 (direct)  273333 (reception)  273418 (Fax) >0 >3 >  >0   ------------------------------  $ Date: Thu, 7 Dec 2000 10:49:34 -0500& From: "Syltrem" <syltrem@videotron.ca>$ Subject: Re: Very weird DCL behavior5 Message-ID: <heOX5.783$He.10598@wagner.videotron.net>w  I I know I use subproprocesses (hey! I'm not that dumb! :-) ). The the real>E question about using one job with subprocesses or many jobs, is: Does F DISMOUNT/NOUNLOAD rewind the tape? - don't get me wrong, rewinding andK unloading are 2 separate functions :-) I thought it did but I may be wrong.t  K If it does not, then I could use SUBMIT with SYNCHRONIZE but... If I submittK 4 backup jobs, I can only synchronize with ONE of them. If this one takes 2 J hours to finish and the other 3 run in 5 minutes, I will loose 3 times oneH hour and 55 minutes of tape drive availability because my main procedureJ will "unfreeze" only after that one very long backp is finished.. No good.9 That makes me loose almost 6 hours of usable backup time.V  F Unless I can implement some sort of AST mechanism in DCL and watch forI completion of any of  the 4 batch jobs (that's exactly what my mailbox isuG doing, by the way) then SYNCH can't do the job. I could use one job per=0 magtape but that's not what what I want, either.  H I think I will end up writing a 15 lines Basic program to handle the I/OG from this damn mailbox. DCL like it or not. It'll be the fastest way of= fixing everything.  F Thanks for all replies (and if someone could answer the question about1 no-unloading/rewinding above, I will appreciate).o   Syltremn      E "Gotfryd Smolik, VMS lists" <gotfryd@stanpol.com.pl> wrote in message D news:Pine.LNX.4.21.0012061439580.10098-100000@irys.stanpol.com.pl...$ > On Tue, 5 Dec 2000, Syltrem wrote: >aC > +The problem with SYNCH is that the backup processes are detacheds	 processesgL > +(I mean, not *real* detached, but separate from my main process). The way Ii >h; >  Probably Hoff describes properly: you use subprocesses ?  >sJ > +do it, I mount the tape in the main procedure, and backups append to itD > +(BACK/NOREW) without having to rewind the tape after every backup command. > +rL > +The SYNCH method would take at least twice as long to complete everythingE > +because of the time required to rewind the tape, then for the next- process > > +to go at logical end-of-tape befors actually doing its job. >m >  Wait a moment.c	 >  Why ??C >t@ >  Althought not ALLOCATing the tape (or braking the allocation)? > is *not* a good idea - if you ask "only" for the "not rewind" = > problem - then say why not simply MOUNT/NOUNLOAD the tape ?y= > (or DISMOUNT/NOUNLOAD, regards how you will handle the tape  > while error) >  > [...] 
 > +Syltrem > [...]y > [...citate from Hoff...] >t? >  Additional request from me: can you answer *after* the citetf. > text ? That is much more readable... Thanks. >oB > +"Hoff Hoffman" <hoffman@xdelta.zko.dec.nospam> wrote in message > [...]lH > +>   I'd probably either use SYNCH and separate batch jobs, or (if you have> > +>   four tape drives), a four-slot batch queue and ALLOCATE mmdu:,mmdu,...L > +>   or ALLOCATE/GENERIC, as this can greatly simplify the synchronization > +>   involved. >  >  Agree. But: >d > [...].G > +>   Looks like something is getting scrozzled deep within DCL, or in F > +>   the code between OpenVMS and the mailbox driver -- DCL can onlyE > +>   use mailboxes with some additional assistance, so you are, um,h? > +>   boldly going where few DCL programmers have gone before.n >nA >  Have use mailboxes from vanilia DCL without problems, even thehC > DCL-created (implementation of the old, known here trick can sendk > anyone who asks) TMPU_ one.m >  >  Will suspect: > 1. as Hoff say:- >-E > +>   Please acquire and apply the current ECOs for whatever OpenVMSe releasea >  - a bug in specific version >e >  or:A > 2. a bug in implementation regard for the SYS$COMMAND/SYS$INPUT3 >  dependence for subprocess.cD >   The "suspected line" looks like b.ex. DCL verify read as command* > or something, and we must remember that: > $ spawn/input=A @B> > (A and B can be both files, mailboxes or other devices) will> > execute commands from B, but when B ends or line starts with! > "$" A can be read for commands.n >o	 >  Check:t > $ spa/inp=tt: @tt:& > %DCL-S-SPAWNED, process GS_1 spawned8 > %DCL-S-ATTACHED, terminal now attached to process GS_1B > $ write sys$output f$environment("depth") ! Executed from /INPUT > 1 , > _$ write sys$output f$environment("depth") > 1c1 > _$  Exit ! Here CTL-Z is send as "end-of-input"5? > $ write sys$output f$environment("depth") ! Executes from "@"t > 0  > $ lo6 >   Process GS_1 logged out at  6-DEC-2000 14:21:49.031 > %DCL-S-RETURNED, control returned to process GSt > $t >g> >  SET VERIFY with the behaviour gives spectacular effects -:)< >  The <NL> is suspected: looks like SEARCH in pipe... ? (or > DCL bug, sure).  > + >  I am little confused with the following:e >r/ > $       Read Backup_Mbx Mbx_Rec /Time_Out=255f >t8 > ...because RMS mailbox read timeout support is limited7 > to /TIME=0 (documented in RMS manual and pointed herec6 > from time to time) - then you must have code without > time-related problems -:)O >  Short check:A >09 > $ crembx tst ! procedure with the TMBU hack on DCL pathl: > %DCL-S-RETURNED, control returned to process MBX204001261 > %DCL-S-RETURNED, control returned to process GS  > $ sh dev tst >j0 > Device                  Device           Error0 >  Name                   Status           Count0 > MBA56:                  Online               0 > $ sh log/full tst H >    "TST" [super,confine] = "_BUK$MBA56" [terminal] (LNM$PROCESS_TABLE) > $ read/time=0 tst line" > %RMS-E-EOF, end of file detected > $ read/time=3 tst line& > [hangs as long the mailbox is empty] >  >  Regards - Gotfryd >  > --G > ===================================================================== H > $ ON F$ERROR("LANGUAGE","ENGLISH","IN_MESSAGE").GT.F$ERROR("NORMAL") - > THEN EXCUSE/OBJECT=MEM0 > $!                        GS@stanpol.zabrze.plG > =====================================================================  >i >o >a   ------------------------------  $ Date: Thu, 7 Dec 2000 10:59:46 -0500& From: "Syltrem" <syltrem@videotron.ca>$ Subject: Re: Very weird DCL behavior5 Message-ID: <SnOX5.785$He.11797@wagner.videotron.net>   J I do this simply because I cannot evaluate how many blocks will have to beF backed up from DISK nnn every day. Sometimes there is a lot, sometimes notihing at all.  L If I say, let's backup disks 1, 2, 3 and 4 to tape drive A. Some days I willH not have enough tapes to backup everything. And some days I will use oneJ cassette from a 10-cassettes loader. And I may not have enough room in the) other loader which backups disks 5, to 8.t  B There's never a chance that I do not have enough tapes, but I must; distribute the files equally to make sure all is backed up.   L So I just append any disk I back up to one set of tapes in a stacker. I haveJ 4 stackers, 4 backup streams. The procedure records where each disk backup* is going, and we are very happy with that.   Syltremr  - <steven.reece@quintiles.com> wrote in messageaB news:OFE38DB61B.A35C0B3B-ON802569AD.0052A490@qedi.quintiles.com... >o >eE > What are you looking to achieve, what do you really require of youre backupsa7 > and what do you have available to do your backups on?e >hG > For our site with multiple tapes in multiple drives (we use a pair of . > TL892s on each production cluster) we have :; > - one batch queue per tape drive with a job limit of one; L > - a batch job that runs each evening (at about tea-time) that submits jobsI > to all of the tape related batch queues in a specified order.  Each job  hasG > a specific release time. >-I > This eliminates the need to do any synchronisation commands as the tape<J > device (say MKA100) is only ever used by a job in the NODE1_MKA100 batchI > queue.  When one job has finished then the next one starts (just like a > > print queue, but using the tape drive as the output device). >.$ > This serves a number of purposes :G > - provided that the tape doesn't overrun then we know precisely whicha > backup is on which tape;K > - each of the backups are only done on one drive so we know which tape iny > which drive/stacker/library;! > - the jobs always run in order;tK > - one tape can contain one or more savesets or can hold part of a saveset ! > if a continuation tape is used;r >eI > Hence, if you need to get back a specific file from your user disk theni you F > can go to your firesafe and get the appropriate user disk backup and > restore the files from that. >-B > One can also play other games like reading in a list file in the
 submissionL > job to run a specific series of jobs on a particular night of the week/dayJ > of the month yet still submit the same batch job to run during the early
 > evening. >mD > Using multiple tapes and just going to the next tape device that'sJ > available starts to suggest something like Veritas as a backup solution, in- > which case one would have to ask WHY???????  >a > Steve. >.. > Syltrem (syltrem at videotron dot ca) asked:J > >>>I use a DCL procedure (what else!) to do my nightly backups. I have 4 > tapeD > drives, and I backup parts of my data (that could be one disk, oneD > directory, one oracle db - but it doesn`t really matter). The main > procedure D > has a list of things to do which it dispatches (via a spawn/nowait command)# > to the next available tape drive.  >1G > When one backup is finiished, the spawned procedure notifies the main K > process that it has finished so that another job can be sent to that tapee > drive (am I clear enough?).r >gK > Upon reading the mailbox, the main procedure sometimes (once every 100 oro0 > 200 times it does the READ command) hits error, > %RMS-E-RSA, record stream currently activeF > when (or right after) reading the mailbox. I do not know why this isK > happening, and the weird thing about this is that the .LOG shows commandsi+ > that do not exist in the .COM file, like:  >e > $       Set NoOn/ > $       Read Backup_Mbx Mbx_Rec /Time_Out=255i= > $       If .not. $(Backup_Status<NL><NL>0,Mbx_Rec)<NL>_Diskf, > %RMS-E-RSA, record stream currently active > $ Error_Trap:R > $       Exit_Status = $STATUS  >t > Note:oL > a) The 2nd line in this .LOG excerpt does not belong to the .COM. Note the# > <NL> characters in the line, too. K > b) The error message does not follow immediately the READ command, but it- > sure is related to it7H > c) The SET NOON seems to have no effect as the procedure jumps into my > erroru > trap.. >m( > The .COM contains the following lines: > $       Set NoOn/ > $       Read Backup_Mbx Mbx_Rec /Time_Out=255s4 > $       If .not. $STATUS then Goto Backup_One_Disk > $       Set On$ > $       Unit = f$extr(0,1,Mbx_Rec). > $       Backup_Status = f$extr(1,10,Mbx_Rec) > E > It seems like the DLC processor has lost its pointer in reading the  commandv > file.U >4 > Can someone help?<<< >a   ------------------------------   Date: 7 Dec 2000 11:39:10 -0500g* From: young_r@eisner.decus.org (Rob Young)$ Subject: Re: Very weird DCL behavior+ Message-ID: <yJHEn8xpzLK3@eisner.decus.org>m  H 	Submit your batch jobs to EXECUTION queues with a JOB LIMIT of 1.  Only7 	1 will be active at a time and the others wait, right?0  - 	Create a seperate queue for each tape drive:g  
 	DRIVE1_BATCHl 	DRIVE2_BATCH , etc.  @ 	These are fed from a generic queue so as one job completes, theD 	next spools off to an execution queue.  The DCL to do the backup is< 	generic enough as it knows which tape drive to backup to by 	parsing the queue name.  A 	Assumption is it doesn't matter which backup goes to which tape a" 	(i.e. you want to get them done).  @ 	You can use a .DAT file to have all the volumes to be backed up? 	in there.  Or you may just be backing up files or somewhere in ; 	between.  We had our most critical volumes in there first.   E 	And yes, I did a DISMOUNT/NOUNLOAD .. regarding rewinding... I don'teE 	think so.  If it was rewinding, it didn't matter to me as my backupso? 	were done in the morning :).  However, looking into this I did.C 	put a SET MAGTAPE command to make sure it was EOF (peeling throughp@ 	a few DSN articles now I am unclear as to my reasoning but I doB 	know we did test restores and did do BACKUP/LIST , etc. from time@ 	to time).  Actually.. re-reading some of the DSN articles shows 	my reasoning was faulty:   G When a new saveset is added to a tape containing multiple savesets, the=7 new saveset will have consecutive TM's, indicating EOV.-  ? 	So my SET MAGTAPE command was overkill.  The question then is,m1 	does a DISMOUNT/NOUNLOAD go to BOT?  Don't know.c     $ !s $       gosub build_commands $ !D $ !4' $       proc_name = "BCK ''disk_to_do'"l' $       set process/name="''proc_name'"l $ !u $ !     Perform the Backup $ !t $       write sys$output ""aP $       write sys$output "Beginning backup of ''disk_to_do' to ''tape_drive' at  ''f$time()'"/ $       on severe_error then call error_handlercE $       backup_command 'disk_path' 'tape_drive':'disk_to_do'.bck/save-K $       write sys$output "Completed backup of ''disk_to_do' at ''f$time()'"A $       write sys$output ""s $ !p $ !a- $       set magtape/end_of_file 'tape_drive': ' $       dismount/nounload 'tape_drive':e7 $       open/write mailtmp 'this_dir'mailtmp.'this_pid'o $       write mailtmp "". $       write mailtmp " Backup command used: " $       write mailtmp ""# $       write mailtmp " ''piece1' "o# $       write mailtmp " ''piece2' "h# $       write mailtmp " ''piece3' "n# $       write mailtmp " ''piece4' "y" $       write mailtmp " ''quals' " $       write mailtmp "" $       close/nolog mailtmp K $       mail/sub="''DISK_TO_DO' ''TAPE_DRIVE' ''LABEL_NAME' ''DISK_PATH'" -o1         'this_dir'mailtmp.'this_pid'  backup_logsr $       delete = "delete"(, $       delete 'this_dir'mailtmp.'this_pid'; $       exit     $ build_command: $       quals = "" $ !  $ !l5 $       if backup_method - "INCR" .eqs. backup_method- $       then( $               quals = quals + "/IMAGE", $               disk_path = "''disk_to_do':" $       else4 $               quals = quals + "/SINCE=BACKUP/FAST"7 $               disk_path = "''disk_to_do':[*...]*.*;*" 
 $       endif1 $ !0 $ !95 $       if backup_method - "RECO" .nes. backup_method  $       then) $               quals = quals + "/RECORD"f
 $       endifa $ !l $ !i6 $       piece1 = "BACKUP/NOREWIND/LABEL=''label_name'"1 $       piece2 = "/IGNORE=(INTERLOCK,LABEL_PROC)" H $       piece3 = "/JOURNAL=TS$BUL:''YYYYMMDD'_''DISK_TO_DO'_''TAPE'.BJL"$ $       piece4 = "/BLOCK_SIZE=65535". $       piece5 = "/MEDIA_FORMAT=''COMPACTION'"K $       backup_command = piece1 + piece2 + piece3 + piece4 + piece5 + quals- $       return     				Robn    ^ In article <heOX5.783$He.10598@wagner.videotron.net>, "Syltrem" <syltrem@videotron.ca> writes:K > I know I use subproprocesses (hey! I'm not that dumb! :-) ). The the real G > question about using one job with subprocesses or many jobs, is: DoesoH > DISMOUNT/NOUNLOAD rewind the tape? - don't get me wrong, rewinding andM > unloading are 2 separate functions :-) I thought it did but I may be wrong.  > M > If it does not, then I could use SUBMIT with SYNCHRONIZE but... If I submitvM > 4 backup jobs, I can only synchronize with ONE of them. If this one takes 2eL > hours to finish and the other 3 run in 5 minutes, I will loose 3 times oneJ > hour and 55 minutes of tape drive availability because my main procedureL > will "unfreeze" only after that one very long backp is finished.. No good.; > That makes me loose almost 6 hours of usable backup time.n > H > Unless I can implement some sort of AST mechanism in DCL and watch forK > completion of any of  the 4 batch jobs (that's exactly what my mailbox issI > doing, by the way) then SYNCH can't do the job. I could use one job per 2 > magtape but that's not what what I want, either. > J > I think I will end up writing a 15 lines Basic program to handle the I/OI > from this damn mailbox. DCL like it or not. It'll be the fastest way ofs > fixing everything. > H > Thanks for all replies (and if someone could answer the question about3 > no-unloading/rewinding above, I will appreciate).0 > 	 > Syltrem- >    ------------------------------  % Date: Thu, 07 Dec 2000 09:50:16 -0800l From: IOlson@dairyworld.comR$ Subject: RE: Very weird DCL behaviorM Message-ID: <763C579A82F7D3118EE400D0B74723D10166B62E@exchsrv.dairyworld.com>e  L Ok, this doesn't address the "rewind" question, but I thought I'd throw this out there anyway.jA Here's an alternative technique I've used in similar situtations:r2 1. spawn off a subprocess, give it a specific nameF 2. go into a wait loop and keep checking if the process is still alive0 3. when it disappears, you can spawn another oneC 4. if you want to know what happened to it (completed vs died, eg),n@    have the subprocess put some status values into the job table  
 Ingemar Olsono* IT Department, Sperling     (604) 444-7367 Dairyworld FOODS     -----Original Message-----+ From: Syltrem [mailto:syltrem@videotron.ca]e Sent: December 7, 2000 7:50o To: Info-VAX@Mvb.Saic.Come$ Subject: Re: Very weird DCL behavior   <snip>   ------------------------------  / Date: Thu, 07 Dec 2000 13:26:13 +0100 (MET_DST)t From: system@msiw60.msi.se1 Subject: Why do I have to run DECW$UTILS:XREFRESH * Message-ID: <00120713261318@msiw60.msi.se>   Hi everyone,9 does any of you have a solution to the following problem.BD I am using a workstation running OpenVMS 7.2-1 and  DECwindows Motif Version 1.2-5 for OpenVMS.B Every time when I resume my session after a "Session Pause" I haveL to run DECW$UTILS:XREFRESH in order to restore background and border colors.K Also when running for example Netscape I have to run DECW$UTILS:XREFRESH in>J another window. This I have to do for every page I want to see or the textG from the previosly displayed pages will be visible. I suspect that somet> DECW$-parameter have the wrong value. Gratefull for any hints.   Carl Gunnar Lindin Manne Siegbahn Laboratorym Stockholm Sweden   ------------------------------  + Date: Thu, 07 Dec 2000 08:34:03 +0100 (MET)s& From: Rudolf Wingert <win@fom.fgan.de>" Subject: XP1000 hardware questions6 Message-ID: <200012070730.IAA15110@sinet1.fom.fgan.de>   Hello,  D we have an XP1000 professionel workstation (667Mhz). The SHOW DEVICEE console program shows me two SCSI adapters. Only the second one (PCI)hD will be used for the internal SCSI devices. But there is also an on-H board SCSI adapter. Does anybody know about this adpater (type of SCSI)?6 Is this adpater useable for SCSI device under OpenVMS?  E The keyboard driver of the DPWS500au is bad, but an keyboard extender D does function normal. This is better then the keyboard driver of theB XP1000. We would like to store the station in a protected computerE room with firealarm. The keyboard, monitor and mouse should be storedtE in the office of the user. By normal workstations (AS0600, DS20E ...)eB we can do this with a limit far away 50 meters. the XP1000 not so.C After 18 meters is finish ("keyboard not plugged in" message). WithpB the keyboard extender (KE101) we can arrive a length of 30 meters.H Other with a PC: 27 meter between PC and Keyboard extender and 36 metersD between extender and keyboard. Is there any extender out, which willA allow to store the XP1000 keyboard more then 30 meters away? Will E a keyboard ghost a solution (simulates a connected keyboard)? Because E if we boot with a short keyboardkabel and change them to a longer, it 
 functions.   TIA and regards Rudolf Wingert   ------------------------------  # Date: Thu, 07 Dec 2000 14:06:29 GMTa* From: Doran Werling <rwscsinc@my-deja.com>& Subject: Re: XP1000 hardware questions) Message-ID: <90o5h1$c5g$1@nnrp1.deja.com>o  6 In article <200012070730.IAA15110@sinet1.fom.fgan.de>,)   Rudolf Wingert <win@fom.fgan.de> wrote:y > Hello, > F > we have an XP1000 professionel workstation (667Mhz). The SHOW DEVICEG > console program shows me two SCSI adapters. Only the second one (PCI)iF > will be used for the internal SCSI devices. But there is also an on-J > board SCSI adapter. Does anybody know about this adpater (type of SCSI)?8 > Is this adpater useable for SCSI device under OpenVMS? > G > The keyboard driver of the DPWS500au is bad, but an keyboard extender F > does function normal. This is better then the keyboard driver of theD > XP1000. We would like to store the station in a protected computerG > room with firealarm. The keyboard, monitor and mouse should be storedcG > in the office of the user. By normal workstations (AS0600, DS20E ...)PD > we can do this with a limit far away 50 meters. the XP1000 not so.E > After 18 meters is finish ("keyboard not plugged in" message). WithrD > the keyboard extender (KE101) we can arrive a length of 30 meters.J > Other with a PC: 27 meter between PC and Keyboard extender and 36 metersF > between extender and keyboard. Is there any extender out, which willC > allow to store the XP1000 keyboard more then 30 meters away? Will.G > a keyboard ghost a solution (simulates a connected keyboard)? BecausegG > if we boot with a short keyboardkabel and change them to a longer, ite > functions. >  >eG All XP1000 systems include an onboard UltraSCSI controller for internaldC disks only. The 667MHz systems include an additional PCI-based WideeF Ultra-2 SCSI (LVD) adapter (SN-KZPCA-AA, factory installed) to improve internal disk performance.  E We are using keyboard extenders on some of our 30+ XP900  systems butSI nothing over 5 meters. You might talk to compaq service about a solution.    Regards,
 Doran Werling  RW/SCS Inc.N    & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------   End of INFO-VAX 2000.683 ************************