1 INFO-VAX	Wed, 05 Jul 2000	Volume 2000 : Issue 373       Contents: Application Future ? Re: Application Future ?! Re: CD burners for OpenVMS Alpha?  DCPS queues. DCPS queues. Re: DCPS queues. Re: DE500 on Alpha VMS 6.2< Re: Directory Sizes (was: good news (for me,  I think) . . .7 Re: Disk I/O Performance (was Re: OpenVMS loses big...) 7 Re: Disk I/O Performance (was Re: OpenVMS loses big...) 7 Re: Disk I/O Performance (was Re: OpenVMS loses big...) 7 Re: Disk I/O Performance (was Re: OpenVMS loses big...) 7 Re: Disk I/O Performance (was Re: OpenVMS loses big...) 7 Re: Disk I/O Performance (was Re: OpenVMS loses big...) 
 Fragmentation  Re: Fragmentation  Re: Fragmentation  Re: Fragmentation ) Re: highwater marking, speed vs. security  How to break up a LAVC ??  Re: How to break up a LAVC ??  Re: LA36 Escape codes  Re: LAVC Cluster Re: LAVC Cluster Re: LAVC Cluster Re: Locate the hotfiles  RE: Locate the hotfiles  RE: Locate the hotfiles  Re: Locate the hotfiles 0 Missing pseudocolor visual (Elsa Gloria Synergy)# Multilingual System Dictionary(ies) ' Re: Multilingual System Dictionary(ies) $ Netscape Fasttrack for OpenVMS 7.2-1( Re: Netscape Fasttrack for OpenVMS 7.2-1( Re: Netscape Fasttrack for OpenVMS 7.2-12 Re: OpenVMS loses big, was:  RE: Compaq advertises2 Re: OpenVMS loses big, was:  RE: Compaq advertises2 Re: OpenVMS loses big, was:  RE: Compaq advertises2 Re: OpenVMS loses big, was:  RE: Compaq advertises Problem with VFYMAIL utility. Problems with the DIRECTORY command (it loops)2 Re: Problems with the DIRECTORY command (it loops) Re[2]: VMS Printing using PCL6% Re: RMS indexed file instatly growing + Re: RMS tuning breaking programs, any ideas ! SCSI hard drive 4 VAX 4700 needed 1 Re: Selling software to corporation AND hobbyists 1 Re: Selling software to corporation AND hobbyists 1 Re: Selling software to corporation AND hobbyists  Shareware: txt2pdf 3.6 Re: Traceback info from AST ?  Trapping a STOP/ID Re: Trapping a STOP/ID Re: Trapping a STOP/ID Re: Trapping a STOP/ID Re: Trapping a STOP/ID Re: Trapping a STOP/ID- Upgrade to VMS 7.2 fails on Help Message File  Re: VAX on Intel?   RE: VMS can tail -f ? if so how? Re: VMS Printing using PCL6  What's wrong with F$LOCATE ?  RE: What's wrong with F$LOCATE ?  Re: What's wrong with F$LOCATE ?  Re: What's wrong with F$LOCATE ?  F ----------------------------------------------------------------------   Date: 5 Jul 2000 11:17:11 +0200 * From: eplan@kapsch.net (Peter LANGSTOEGER) Subject: Application Future ? * Message-ID: <3962fd17$1@news.kapsch.co.at>  7 As we (should) all see now a VMS boom, I'd like to ask: ! Do the ISVs already see it, too ?   + Will SAP R/3 (or R/4) again deploy on VMS ? J Will EUREX (the european stock exchange) revert the decision against VMS ?K Will ORACLE make use of the VMScluster advantages with the next DB version? G Will ADOBE deliver its products (at least the Acrobat Reader) for VMS ? L Will REAL NETWORKS deliver its products (at least the Real Player) for VMS ?G Will MACROMEDIA deliver its products (at least the Shockwave) for VMS ?   J At least the EUREX decision against VMS (and AIX) is a major pain, becauseH it will force almost all european bank institutes away from VMS over theL not so long run. And this is exactly the high end segment DEC (and Q) wantedO to see VMS in (only). If this decision (done, because of eg. missing JAVA V1.2) . will not change back, then IMHO goodbye VMS...  N And don't tell me, that this is "only the EUREX client, not the EUREX server".( It _will_ force the banks away of VMS...   --  < Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111-2651; Network and OpenVMS system manager  Fax.    +43 1 81111-888 < FBFV/Information Services           E-mail  eplan@kapsch.netF <<< KAPSCH AG  Wagenseilgasse 1     PSImail PSI%(0232)281001141::EPLANH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"N "VMS is today what Microsoft wants Windows NT V8.0 to be!" Compaq, 22-Sep-1998   ------------------------------  % Date: Wed, 05 Jul 2000 13:25:18 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> ! Subject: Re: Application Future ? , Message-ID: <39636F7B.757F0AC8@videotron.ca>   Peter LANGSTOEGER wrote:I > Will MACROMEDIA deliver its products (at least the Shockwave) for VMS ?   K Flash players and SDKs are available. The VMS group, or anyone could make a N port if they wanted to. It would be far easier for Compaq to do the port sinceL it has the money to do so. If an individual does the port, he has no hope of
 making money.    ------------------------------  , Date: Wed, 05 Jul 2000 18:18:40 +0200 (CEST): From: "Gotfryd Smolik, VMS lists" <gotfryd@stanpol.com.pl>* Subject: Re: CD burners for OpenVMS Alpha?I Message-ID: <Pine.LNX.4.21.0007051807570.8705-100000@irys.stanpol.com.pl>   8 On Mon, 3 Jul 2000, Brian Schenkenberger, VAXman- wrote: [...my:]= +> Oh, yes.... Have found one old message in the "in progess"  [...] ; +>On Mon, 17 Jan 2000, Brian Schenkenberger, VAXman- wrote:  [...looong...]- +Gee, Poland has some really slow technology!   :  Yes, I am aware of that... -;) That was *my* (personally) mail technology -:)   I +FYI, even with a second *NEW* disk installed in the toy box, Weendoze'98 H +would not install.  It saw the Weendoze NT disk and refused to install.F +The disk was physically disconnected to get it to install.  There is G +still some braindamage as Weendoze'98 insists that D: is a CDrom when,  +in fact, F: is the CD. :(  A  Ok. What have pointed before: the hammer to get "roughly proper" 7 work with Win9X is "disable the unneeded disk in BIOS". =  Then Win9X and WNT have control of the "device letter" name, 2 you *can* assign any name to the disk if you want.:  Also you *can* name the CD-ROM as F:, G: or what you want> (my standard setting: R: is CD-ROM or DVD, W: is CD-writer !).B  If you interset in the point(s) for W9X then send me private mailD to Reply-To:, I don't want be flamed because Not-The-Group subject !  + +But, whaddaya want from monopoly$chlock???   :  Unfortunatelly the problem has reason: DOS compatibility.;  For that reason have little defend Microsoft; rest of your   "blue screen" points are right !    Regards - Gotfryd   --  E ===================================================================== F $ ON F$ERROR("LANGUAGE","ENGLISH","IN_MESSAGE").GT.F$ERROR("NORMAL") - 		THEN EXCUSE/OBJECT=ME . $!                        GS@stanpol.zabrze.plE =====================================================================    ------------------------------  % Date: Wed, 05 Jul 2000 18:42:59 +0010 % From: paddy.o'brien@zzz.tg.nsw.gov.au  Subject: DCPS queues. 5 Message-ID: <01JRF2BKUAN6001FWA@tgmail.tg.nsw.gov.au>   I I have a new ES40.  VMS 7.2 and to my knowledge the latest DCPS software.   Q I have DECnet IV, so I look at my data base (MCR NCP) and find that queue P1106l  / is in both the permanent and current databases.   & The code I use to set up the queue is:  B $!	Initialise and start the execution queue DCPS$'P1' on node 'P1'J $!	and the associated generic queues for single and double sided printing. $ 
 $ SET NOON $ 8 $!	Define a logical for ALL device control libries used.- $!	Define a logical for double sided printing  $ J $ DEFINE/SYSTEM/EXEC/NOLOG DCPS_LIB DCPS$DEVCTL,"SYSDEVCTL/DATA_TYPE=ANSI"; $ DEFINE/SYSTEM/EXEC/NOLOG DCPS$'P2'_2S_PARAMETER "SIDES=2"  $ ' $! Initialize execution queue DCPS$'P1'  $  $ INITIALIZE	/QUEUE-	 		/START-  		/ON='P1' -   		/LIBRARY=DCPS_LIB-"                 /NOENABLE_GENERIC-=   		/DEFAULT=(FORM=DCPS$NEW_DEFAULT,NOFEED,NOFLAG,NOTRAILER)-    		/SCHEDULE=NOSIZE-  		/PROCESSOR=DCPS$SMB- 		/FORMS=DCPS$NEW_DEFAULT- 		DCPS$'P1'  $ ; $! Initialize and start the generic queues 'P2' and 'P2'_2S  $  $ INITIALIZE /QUEUE -               /START - #              /GENERIC = DCPS$'P1' -               'P2'  $  $ INITIALIZE /QUEUE -               /START - #              /GENERIC = DCPS$'P1' -               'P2'_2S $  $ EXIT    # So, @START_DCPS_QUEUE P1106l Q1154A   M However, unlike my other DCPS queues, when I submit a job to Q1154A, it only  ! shows "starting"  the queue busy.   N Nothing ever gets printed.  What have I omitted?  BTW, the forms library also  exists.    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, ; but that little whizz-bang apostrophe gives me little spam.    ------------------------------  % Date: Wed, 05 Jul 2000 06:39:33 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com> Subject: DCPS queues. 7 Message-ID: <200007050639_MC2-AB23-D949@compuserve.com>            Paddy,  =         What is "queue P1106I" doing in your DECnet database?   H         If *node* P1106I is not part of the local cluster, I don't think this is going to work!  @ Message text written by INTERNET:paddy.o'brien@zzz.tg.nsw.gov.auJ >I have a new ES40.  VMS 7.2 and to my knowledge the latest DCPS software= .   J I have DECnet IV, so I look at my data base (MCR NCP) and find that queue=   P1106l =  / is in both the permanent and current databases.   & The code I use to set up the queue is:  G $!      Initialise and start the execution queue DCPS$'P1' on node 'P1' E $!      and the associated generic queues for single and double sided 	 printing.  $ 
 $ SET NOON $ = $!      Define a logical for ALL device control libries used. 2 $!      Define a logical for double sided printing $ J $ DEFINE/SYSTEM/EXEC/NOLOG DCPS_LIB DCPS$DEVCTL,"SYSDEVCTL/DATA_TYPE=3DAN= SI" = $ DEFINE/SYSTEM/EXEC/NOLOG DCPS$'P2'_2S_PARAMETER "SIDES=3D2"  $ ' $! Initialize execution queue DCPS$'P1'  $  $ INITIALIZE    /QUEUE-                  /START-                  /ON=3D'P1' -$                 /LIBRARY=3DDCPS_LIB-"                 /NOENABLE_GENERIC-J                 /DEFAULT=3D(FORM=3DDCPS$NEW_DEFAULT,NOFEED,NOFLAG,NOTRAIL= ER)-#                 /SCHEDULE=3DNOSIZE- &                 /PROCESSOR=3DDCPS$SMB-*                 /FORMS=3DDCPS$NEW_DEFAULT-                 DCPS$'P1'  $ ; $! Initialize and start the generic queues 'P2' and 'P2'_2S  $  $ INITIALIZE /QUEUE -               /START - %              /GENERIC =3D DCPS$'P1' -               'P2'  $  $ INITIALIZE /QUEUE -               /START - %              /GENERIC =3D DCPS$'P1' -               'P2'_2S $ =    $ EXIT    # So, @START_DCPS_QUEUE P1106l Q1154A   G However, unlike my other DCPS queues, when I submit a job to Q1154A, it  only =  ! shows "starting"  the queue busy.   H Nothing ever gets printed.  What have I omitted?  BTW, the forms library also =   exists.    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  E Either "\'" or "\s" (to escape the apostrophe) seems to work for most  people, < but that little whizz-bang apostrophe gives me little spam.<   ------------------------------  # Date: Wed, 05 Jul 2000 15:30:42 GMT + From: Paul Anderson <panderson@genicom.com>  Subject: Re: DCPS queues. C Message-ID: <panderson-3B2F63.11305905072000@news.bellatlantic.net>   6 In article <01JRF2BKUAN6001FWA@tgmail.tg.nsw.gov.au>, & paddy.o'brien@zzz.tg.nsw.gov.au wrote:   > $ INITIALIZE /QUEUE-
 >     /START-  >     /ON='P1' - >        /LIBRARY=DCPS_LIB- $ >                 /NOENABLE_GENERIC-B >        /DEFAULT=(FORM=DCPS$NEW_DEFAULT,NOFEED,NOFLAG,NOTRAILER)- >        /SCHEDULE=NOSIZE- >     /PROCESSOR=DCPS$SMB- >     /FORMS=DCPS$NEW_DEFAULT- >     DCPS$'P1'   G What type of printer are you using?  What you describe would work only  % for a PrintServer printer via DECnet.   = > $ DEFINE/SYSTEM/EXEC/NOLOG DCPS$'P2'_2S_PARAMETER "SIDES=2"   H FYI, you are not using the recommended DCPS$STARTUP.COM as a way to set > up queues.  You may be missing some queue setup details since G DCPS$STARTUP.COM sets up logical names depending on queue features you  I request, as well as handle long printer name or address specifications.   D These logical names might be overlooked by using your own procedure.  J In any case, looking at the comments in DCPS$STARTUP.COM might help you.  F Also enable operator messages with a REPLY /ENABLE and see what error * messages appear when you start your queue.   Paul   --  "    Paul Anderson, DCPS Engineering"    GENICOM Corporation, Gardner MA   ------------------------------  % Date: Tue, 04 Jul 2000 19:22:55 -0700 4 From: Roger Wiechman <Roger_Wiechman@HMC.nospam.Edu># Subject: Re: DE500 on Alpha VMS 6.2 . Message-ID: <39629BFE.8BD336AA@HMC.nospam.Edu>   "John E. Malmberg" wrote:  > / > Main, Kerry <Kerry.Main@COMPAQ.COMpaq> wrote:  > To: <Info-VAX@Mvb.Saic.Com> % > Sent: Friday, June 30, 2000 7:34 AM % > Subject: RE: DE500 on Alpha VMS 6.2  > 	 > > John,  > > H > > THe DE500-BA will work on OpenVMS V6.2 with the appropriate patches.	 > Contact D > > the CSC or DSNlink (if available) as there are articles on this. > > J > > A big issue which I constantly run into is the issue of auto-negotiateN > > feature. Depending on the age of equipment involved, there does not appearK > > to be full compatibility within the so called standards for this (or at > > > least in its implementation with older network equipment). > > M > > As a suggestion, I recommend physical settng the host ethernet adapter to J > > 100Mb, Full Duplex AND the apprpriate network switch port to the same. > K > Do *not* arbitrarily set the Duplex setting based on a newsgroup posting.   = * Probably a good suggestion.  Always consider the source. :)   A   To keep "on thread", I'd add that I've found that the DE500 was E   very poor at autonegotiation with our 3COM switches.  So I'd second F   the vote on just setting it to match the connection point.  In fact,    I'd set the other end as well.   > N > The setting must match your network environment.  As Kerry points out, there5 > are problems when the auto-negotiate guesses wrong.    L > In most of the cases that I have read about, and the few that I have seen,N > the condition that has occurred is that full duplex was negotiated, and that# > caused excessive network chatter.   E * Actually, what happens is collisions that should not have been.  If  the full duplex F  side wants to send a packet, it would ignore any incoming data and go	 ahead and ?  send it.  The half duplex side would then immediatly declare a  collision, forcingH  resending.  In a busy network, full or half duplex, this can be deadly.  8 > The DUPLEX setting must match the network requirement.  * * A somewhat broad, but correct statement.   M > In the environment I am in, it *MUST* be half duplex, if it is connected to  > the main network.    * Time to buy some switches.   D > When the cards are set to full duplex, it assumes that there is noE > collisions on the network because all devices are connected through N > switching hubs.  That is you have no thin-wire, thick-wire, or non-switching > hubs on your network.   H * While true in the general sense, most environment have some switches.  Thick wire, A   thin-wire, and twisted pair HUBs -CAN- coexist as long as their  collision domain!   is isolated from other systems.   G   We routinely have 100Mb full duplex devices on a switch that also has  a 10MbF   half duplex port serving a 10Mb hub.  The switched port isolates the hub's C   collision domain from the rest of the network and the full duplex  devices.     Roger at HMC dot EDU
   (wa6zvp)$   Harvey Mudd College, Claremont, CA   ------------------------------    Date: 05 Jul 2000 09:54:21 +0200G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de> E Subject: Re: Directory Sizes (was: good news (for me,  I think) . . . H Message-ID: <y47lb1xb9u.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  9 "David J. Dachtera" <djesys.nospam@earthlink.net> writes:   F > Very true! ...and for those OpenVMS and VAX/VMS versions which don'tI > support CREATE/DIRECTORY/ALLCOATION=nnnn, I have a DCL proc. which will B > pre-allocate directories so they are not (seriously) fragmented.  > 'scuse me - don't ODS2 directory files have to be contiguous!?  D Of course, pre-allocation is still a good idea in certain scenarios.   	Jan   ------------------------------  / Date: Wed, 05 Jul 2000 09:07:57 +0200 (MET DST)0& From: Rudolf Wingert <win@fom.fgan.de>@ Subject: Re: Disk I/O Performance (was Re: OpenVMS loses big...)/ Message-ID: <200007050614.IAA25675@fom.fgan.de>r   Hello,  C our expirience is, that the read performance of OpenVMS is not bad.M@ We have tested that with three OSes. WindowsNT, Sun Solaris 2.6,D OpenVMS 7.1-1H1. The disk was a 18.2GB IBM DGHS disk. We did measure@ the read into the null device. On all three OSes we got a resultC nearly the same (14.9MB/s...15.2MB/s). Other the write performance. F In case of disabled disk drive write cache the performance growth downF to 3.6MB/s under OpenVMS. On the other OSses we did see a little lowerE performace (write a 128MB file) then the read performance (14.5MB/s).tD I think this was the handling time (read from one disk, write to theC other). Bringing the OpenVMS disk in an unsecure state (write cachetB enabled with the freeware tool RZDISK) growth up the performace toC the same value as the other OSses (14.5MB/s for an 128MB and 1024MBeB file). The problem is, that the user can't decision (normaly) that. the write cache should be enabled or disabled.   Regards Rudolf Wingert   ------------------------------    Date: 05 Jul 2000 10:51:22 +0200G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>s@ Subject: Re: Disk I/O Performance (was Re: OpenVMS loses big...)H Message-ID: <y4vgylvu2d.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  C So have you tried using QIOs asynchronously, or FAST_IO? Or are youCG complaining that the C RTL read() doesn't give you stellar performance?    	Jan   ------------------------------  # Date: Wed, 05 Jul 2000 14:16:06 GMTa From: kosh4711@my-deja.com@ Subject: Re: Disk I/O Performance (was Re: OpenVMS loses big...)) Message-ID: <8jvfv7$3q7$1@nnrp2.deja.com>   ( In article <8jt8qo$ffg$1@pyrite.mv.net>,*   "Bill Todd" <billtodd@foo.mv.com> wrote: > , > Doug W. <dashw459@aol.comeatspam> wrote in message ( > news:20000704104302.29762.00000208@ng- mb1.aol.com... >g > ...P > * > > While the changes you referred to will benefit all customers, VMS and > COMPAQ. > > storage products have inherent problems no amount of RMS or cache tweaking- > will > > resolve. >- > ...- >- >   Frankly, the aging RAID / > > solutions available for VMS just plain suck) on both cost and performance > when1 > > compared to competitors.  ULTRA 3 is not eveno supported.  Tacking fiber in* > > front of old products does not inspire confidence.   Too many COMPAQ4
 > products1 > > are UNIX or PC only.  Even if they work underg VMS (proliant hardware),
 > they are* > > not qualified.   The VMS low cost RAID hardware (KZP) needed serious 
 > redesign- > > years ago.  A  KZP trying to spiral is and engineering embarassment.   TheS > HSZr. > > does somewhat better,  but is ridiculously expensive.  The salesi > literature is(0 > > confusing at best (ULTRA2 like rates?).  The sales people just don't know > and 1 > > recommend products that can't cut it.   Theree are no programs to exercisef > and / > > display Max I/O (shame on engineering, looke whats available on the PC).  > The . > > software RAID can't get out of its own way and constantly makes an 
 > observer. > > question why it is doing what it is doing. Binding 2 HSZs makes a good  > case# > > for opensource and peer review.. >t/ > Storage has been one of Compaq's (and earliere DEC's) strengths for quite a- > while, and if it's slipping that's serious.M People in comp.os.vms don't / > seem as sensitive to high storage performance  as some (witness the reactionA0 > to the caching discussion), so this may not be the best place to gain/ > corroboration for your assessment or generate  the pressure on Compaq that / > your description of the situation suggests isa
 necessary. > 0 > Corroboration comes first:  if Compaq can sell all the storage it wants to,+ > then it'll take a good deal more than oneu critical customer to make it+ > listen.  But I already know of a bunch ofc" reasons that large storage vendors. > like Compaq and EMC ought to be looking over their shoulders and trying to / > stay technically current, and this is another  big one if the situation is , > anywhere nearly as bad as you describe it. >  > ...h >o- > > At the bottom of the screen you will findv UCB$L_MAXBCNT.    It will read > FE00.a1 > > Thats the max number of bytes you can xfer toa the device, 127 blocks.  If  > youC+ > > want to do 80MB/sec too many splits arer" required.  If I seem angry keep in > mind, > > that I have experienced many jokes about offloading disk I/O VMS can nott > handle- > > to a cheap PC running NT.  COMPAQ please,r please fix the '127 block 
 > feature' > > quickly. > >l > >s > >n- > > So what can one do?  COMPAQ RAID hardware0 won't cut it, its software RAIDw > is/ > > pitiful and 3rd party products have troubleo under VMS.  We are buying 2a > disk1 > > controllers (ULTRA 2 ) and 4 drives.  We havel written a small program that > does0 > > disk striping  on a file basis for reads and writes only.  No fancy RMS- > > features just QIOs in buckets of 127.  NoT parity disk, hot swapping,
 > shadows,0 > > backups or anything else.  A hack and kludge surely.  But nothing COMPAQa > makes . > > comes close to it in performance.  And the cost is rock bottom.  We wereh/ > > willing, ready and eager to spend a millions but could not find a product > that1 > > worked.  As for the inevitable comments aboute software kludges, untils > there are - > > VMS products available that offer similart performance there are no good  > > alternatives.s >i' > One should point out that typical IDE!$ controllers (at least historically -/ > it's possible that the situation has improved  recently, but I think that- > would have cropped up in recent discussions  elsewhere if it had) support- > only 16 or 17 memory segments per transfer,t which in the NT environmentv1 > translates to 64 KB worth of 4 KB pages.  So one NT at least IDE drives0 > suffer from the same split-I/O size limit that VMS has.  And while SCSI does=. > not have a similar limitation, NT may handle both the same way (and at leastO1 > some Unix systems default the max disk transfer  size to 128 KB or so, though > that can be bumped up IIRC). >=. > But with SCSI drives, unless the adapter and drive themselves can't handle 1 > the command-processing overhead generated by 64e KB chunks (which seems* > unlikely), it's not clear why using that request size (and building up a + > request queue at the disk to avoid missede" rotations) should do anything more/ > than expend some additional system processingP resources that larger chunks0 > could conserve.  And indeed if you've achieved reasonable performance using- > QIOs, that suggests that this is indeed thec case.u > 1 > So if QIOs are dramatically more effective thani RMS in this situation,0 > either you weren't taking advantage of RMS, or RMS needs serious work.  To an0 > first approximation, RMS block I/O facilities, using asynchronous0 > multi-buffering as presumably you are with the QIOs, should be competitive . > with direct QIOs:  did you try that and find that this is not the case?  Andc, > if you want RMS to handle the asynchronous  multi-buffering for you, you can0 > use its read-ahead and write-behind mechanisms (though IIRC they're limited( > to sequential-file use, and that might automatically bring in, > record-processing overheads as well, which could make a noticeableI! > difference):  did you try that?u >', > If indeed you tried these things and still found dramatic differences1 > between RMS and QIOs, then that's important for  Compaq to know about (and 1 > there are people here who will likely make sureC they find out) - so please/ > expand on the description of your experimentsl if you have time to. >k > >o1 > > We are certain COMPAQ does not understand our  needs and wonder where they  > areu. > > going.  We do not understand why customers are left to qualify COMPAQ's0 > > hardware for VMS or write their own striping	 software.m >i, > The specific deficiencies you found in the striping software (at least my1 > vague impression was that VMS supported RAID-0-0 style striping) would likely1 > be of interest to the development group, shouldo you care to report themr > here.p > 1 > My understanding is that Compaq qualifies diskst for VMS mostly to ensure- > that they correctly support multi-initiator: operation (for clustering),B, > which has historically been a problem even! though the SCSI specification has00 > been unambiguous in this area for a long time. Since the typical VMS system- > is clustered, and since most commodity SCSI  drives seem to work with VMS. > just fine in non-clustered operation (though it's been said that poor1 > support for some mode pages has caused problems. in a few cases), this / > attitude is at least somewhat understandable.w >=/ > But it certainly wouldn't be unreasonable fors Compaq to qualify at least its > own storage products on VMS. >a >    We are worried no1 > > other customers are complaining and wonder if! COMPAQ realizes what its > > competitors providing. >o/ > My guess would be that no other VMS customersq are complaining because theree/ > aren't many doing the kinds of things you aree with VMS (which is far fromt) > saying that this market shouldn't be ofl interest to VMS).r >i+ > My other guess would be that Compaq is asC clueless as to what its_- > competitors are providing as it is in otherP areas.  The VMS development20 > group still has a bunch of competent people in it, but they are a bit, > isolated and likely more than a bit shell- shocked as well after so many  > years of neglect.  >  > - bill >N > >L > > Sorry to dump on you,  > > Screaming  not streaming,A > > DW >N >A    . I try not to get involved in these discussions but...  ' Have you tried writing a simple program , doing let's say, 32k transfers using $qio on- HSG80 with writeback caching (and a recent FCh) adapter, the Emulex LP8000 aka KGPSA-CA)?N  $ Hint, > 50mbyte/sec is attainable...  , That doesn't mean there aren't problems with, RMS and/or directory performance (even after/ the 7.2 enhancements.. Certainly a B-tree basedC1 redesign with journaling for metaupdates would be  quite welcome)) but in many mission-critical environmentsy/ (we do custom TP systems, as an example) having,, to use QIO (and or FAST_IO) directly isn't a% problem, we prefer it to using RMS...t   /m    & Sent via Deja.com http://www.deja.com/ Before you buy.a   ------------------------------  $ Date: Wed, 5 Jul 2000 11:29:15 -0400' From: "Bill Todd" <billtodd@foo.mv.com>u@ Subject: Re: Disk I/O Performance (was Re: OpenVMS loses big...)( Message-ID: <8jvk5s$kbq$1@pyrite.mv.net>  K <kosh4711@my-deja.com> wrote in message news:8jvfv7$3q7$1@nnrp2.deja.com...    ...t  0 > I try not to get involved in these discussions > but... >-) > Have you tried writing a simple program1. > doing let's say, 32k transfers using $qio on/ > HSG80 with writeback caching (and a recent FCD+ > adapter, the Emulex LP8000 aka KGPSA-CA)?t >t& > Hint, > 50mbyte/sec is attainable...  I This is not all that impressive a number, given that the last FC externalTG RAID I happen to have been involved with (a Ciprico 8 + 1 RAID-3 array,LG about 2 years ago) would sustain about 85 MBytes/sec without write-backNL caching enabled (not that write-back caching should be all that effective in< enhancing streaming video applications that use asynchronous multi-buffering).p   - bill   ------------------------------  $ Date: Wed, 5 Jul 2000 11:39:51 -0400' From: "Bill Todd" <billtodd@foo.mv.com>N@ Subject: Re: Disk I/O Performance (was Re: OpenVMS loses big...)( Message-ID: <8jvkpl$lm7$1@pyrite.mv.net>  1 Rudolf Wingert <win@fom.fgan.de> wrote in messageu) news:200007050614.IAA25675@fom.fgan.de...l > Hello, >dE > our expirience is, that the read performance of OpenVMS is not bad.eB > We have tested that with three OSes. WindowsNT, Sun Solaris 2.6,F > OpenVMS 7.1-1H1. The disk was a 18.2GB IBM DGHS disk. We did measureB > the read into the null device. On all three OSes we got a resultE > nearly the same (14.9MB/s...15.2MB/s). Other the write performance.yH > In case of disabled disk drive write cache the performance growth downH > to 3.6MB/s under OpenVMS. On the other OSses we did see a little lowerG > performace (write a 128MB file) then the read performance (14.5MB/s).nF > I think this was the handling time (read from one disk, write to theE > other). Bringing the OpenVMS disk in an unsecure state (write cachenD > enabled with the freeware tool RZDISK) growth up the performace toE > the same value as the other OSses (14.5MB/s for an 128MB and 1024MBlD > file). The problem is, that the user can't decision (normaly) that0 > the write cache should be enabled or disabled.  L The normal approach to obtaining good streaming performance from a SCSI diskK is to use asynchronous multi-buffering such that the disk always has one or J more outstanding requests ready to be satisfied, and can just keep sendingD or receiving the data without pause.  This does not require enablingI write-back caching for writes to perform well.  You can ask RMS to handle G this for you (using its read-ahead/write-behind options) in some cases.t  H If you weren't doing this, the reason you obtained decent streaming-readH performance was because the disk was automatically pre-fetching for you.K Depending on the intelligence of its algorithm, you might or might not have:B been able to do somewhat better by doing your own multi-buffering.   - bill   >o > Regards Rudolf Wingert >  >M   ------------------------------  # Date: Wed, 05 Jul 2000 16:46:07 GMTn From: kosh4711@my-deja.com@ Subject: Re: Disk I/O Performance (was Re: OpenVMS loses big...)) Message-ID: <8jvoo9$g0n$1@nnrp1.deja.com>,  ( In article <8jvk5s$kbq$1@pyrite.mv.net>,*   "Bill Todd" <billtodd@foo.mv.com> wrote: >t9 > <kosh4711@my-deja.com> wrote in message news:8jvfv7$3q7  $1@nnrp2.deja.com... >e > ...a >u2 > > I try not to get involved in these discussions
 > > but... > >r+ > > Have you tried writing a simple programt0 > > doing let's say, 32k transfers using $qio on1 > > HSG80 with writeback caching (and a recent FCn- > > adapter, the Emulex LP8000 aka KGPSA-CA)?h > >u( > > Hint, > 50mbyte/sec is attainable... >oB > This is not all that impressive a number, given that the last FC externalB > RAID I happen to have been involved with (a Ciprico 8 + 1 RAID-3 array,E > about 2 years ago) would sustain about 85 MBytes/sec without write-a backA > caching enabled (not that write-back caching should be all thatk effective in> > enhancing streaming video applications that use asynchronous > multi-buffering).e >i > - bill >s >s  > Yes I am aware that there are controllers out there which haveG better performance than the HSG80, however my point was simply that VMS D attained the documented max bandwith for large sequential transfers,B which can be found in some HSG document on the storageworks site).@ 50M/sec is certainly "good enough" for us, other factors such as@ reliability (including multibus failover are much more important in our particular context.      & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  $ Date: Wed, 5 Jul 2000 10:09:11 +0200C From: "Martin Francois" <Francisco.MartinLeon@Cockerill-sambre.com>U Subject: FragmentationE Message-ID: <519A71327421D2118C2000805F65E04D0A9B8C12@DUN-INT-NEWS01>5  - How   I can see the fragmentation of a file ?j   ------------------------------  % Date: Wed, 05 Jul 2000 14:01:28 +0400z4 From: Valentin Likoum <valentin.likoum@ncc.volga.ru> Subject: Re: Fragmentation5 Message-ID: <15821268996.20000705140128@ncc.volga.ru>e  P On 05.07.2000 Martin Francois <Francisco.MartinLeon@Cockerill-sambre.com> wrote:  N MF> Received: from portal.west.saic.com (portal.west.saic.com [198.151.12.15])C MF>         by www.ncc.volga.ru (8.8.7/8.8.7) with SMTP id NAA01381tO MF>         for <val_mail@NCC.VOLGA.RU>; Wed, 5 Jul 2000 13:45:33 +0500 (SAMST)t9 MF>         (envelope-from Info-VAX-Request@Mvb.Saic.Com)i8 MF> Received: from cpmx.saic.com by portal.west.saic.comQ MF>           via smtpd (for [195.239.151.229]) with SMTP; 5 Jul 2000 08:45:29 UT@X MF> Received: from [139.121.95.12] by cpmx.mail.saic.com; Wed, 5 Jul 2000 01:44:20 -0700H MF> Received: from cpmx.mail.saic.com ([139.121.95.10]) by 139.121.95.12: MF>   (Norton AntiVirus for Internet Email Gateways 1.0) ;* MF>   Wed, 05 Jul 2000 08:44:19 0000 (GMT)U MF> Received: from MVB.SAIC.COM by cpmx.mail.saic.com; Wed, 5 Jul 2000 01:41:55 -0700d MF> X-Newsgroups: comp.os.vmsh( MF> Date: Wed, 5 Jul 2000 10:09:11 +0200I MF> Message-Id: <519A71327421D2118C2000805F65E04D0A9B8C12@DUN-INT-NEWS01>iG MF> From: "Martin Francois" <Francisco.MartinLeon@Cockerill-sambre.com>S MF> Subject: Fragmentation MF> Lines: 3 MF> X-Priority: 3r MF> X-MSMail-Priority: Normal : MF> X-Newsreader: Microsoft Outlook Express 5.00.2314.1300< MF> X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300 MF> To: Info-VAX@Mvb.Saic.Comr! MF> X-Gateway-Source-Info: USENETr, MF> X-UIDL: f4e9980c182c941f8ac86195563c3455  1 MF> How   I can see the fragmentation of a file ?B  3   $ dump /header /block=count=0  my_fragmented_filet    and check "Map area" section   --     Valentin Likoumr   valentin.likoum@ncc.volga.ru   ------------------------------  , Date: Wed, 05 Jul 2000 11:52:08 +0200 (CEST): From: "Gotfryd Smolik, VMS lists" <gotfryd@stanpol.com.pl> Subject: Re: FragmentationI Message-ID: <Pine.LNX.4.21.0007051147100.7279-100000@irys.stanpol.com.pl>a  * On Wed, 5 Jul 2000, Martin Francois wrote:  . +How   I can see the fragmentation of a file ?  #  DUMP/HEADER/BLOCK=COUNT=0 filenamee;  and check for "map area". Any "retrival pointer" points to ?  contiguos block of disk blocks; if the file is really big then0?  also contiguos space is split for description in more ponters,a8  (because the block count is limited), but you can check=  the size of "hole" as differrence of LBN+couns vs. next LBN.,    Regards - Gotfryd   -- .E =====================================================================gF $ ON F$ERROR("LANGUAGE","ENGLISH","IN_MESSAGE").GT.F$ERROR("NORMAL") - 		THEN EXCUSE/OBJECT=MEe. $!                        GS@stanpol.zabrze.plE =====================================================================t   ------------------------------  % Date: Wed, 05 Jul 2000 14:02:37 +0100l- From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>  Subject: Re: Fragmentation) Message-ID: <396331ED.72727C7E@bbc.co.uk>d   Martin Francois wrote:  / > How   I can see the fragmentation of a file ?n  : If you have DFO (Compaq's Disk File Optimizer) the command is   $ defrag show my$disk/volume  2 I don't know about other defragmentation products.  A There is no easy way that I am aware of to get this info from VMSeB without installing a defragmnentation product or writing your own.   HTH    --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.uks  A I speak for myself only and my views in no way represent those ofs MedAS or the BBC.    ------------------------------    Date: 05 Jul 2000 10:32:21 +0200G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>t2 Subject: Re: highwater marking, speed vs. securityH Message-ID: <y4ya3hvuy2.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  6 "Robert Deininger" <rdeininger@mindspring.com> writes:  G > But recently I caught highwater marking causing a big performance hit'L > on our oldest microvax (with slow disks that can't afford the extra pain).K > The files in question were sequential with fixed-length records, and weresM > created from Fortran using direct access mode (i.e. by record number). Even I > though the file was written almost totally in sequential order, RMS wash@ > definitely doing something to every block on each file extend.  I When the file is extended, RMS must mark all the new records added to the  file as empty.   	Jan   ------------------------------  % Date: Wed, 05 Jul 2000 13:11:46 -0400k- From: "Dyer, Steve J." <Steve.Dyer@alcoa.com>o" Subject: How to break up a LAVC ??I Message-ID: <1FB3C0C8AE67D311991F0000E896198A0ADA3F@na-war.war.alcoa.com>i  K I have a 3 node LAVC composed of a uVax 3100, and two VAX 4000s. All at VMSi 5.4-2.  H I need to make each a standalone system. I couldn't see where the manualI addressed this, as permanently breaking up a cluster is probably unusual.     Could it be as easy as this: ???   On each system: ; 1. edit MODPARAMS and set VAXCLUSTER=0 and CHECK_CLUSTER=0.  2. run  AUTOGEN    Then reboot all 3 systems.   Thanks for any advice.  
 Steve Dyer Steve.Dyer@alcoa.com
 Alcoa,Inc.   ------------------------------   Date: 5 Jul 2000 19:17:31 +0200l* From: eplan@kapsch.net (Peter LANGSTOEGER)& Subject: Re: How to break up a LAVC ??* Message-ID: <39636dab$1@news.kapsch.co.at>  y In article <1FB3C0C8AE67D311991F0000E896198A0ADA3F@na-war.war.alcoa.com>, "Dyer, Steve J." <Steve.Dyer@alcoa.com> writes:kL >I have a 3 node LAVC composed of a uVax 3100, and two VAX 4000s. All at VMS >5.4-2.o    Ouch. That's really old stuff...  I >I need to make each a standalone system. I couldn't see where the manualtJ >addressed this, as permanently breaking up a cluster is probably unusual. > ! >Could it be as easy as this: ???B >n >On each system:< >1. edit MODPARAMS and set VAXCLUSTER=0 and CHECK_CLUSTER=0.   VAXCLUSTER=0 is enough   >2. run  AUTOGEN >e >Then reboot all 3 systems.   O Yes. But if you got common files like: SYSUAF, RIGHTSLIST, NETPROXY, NET$PROXY,eI QMAN$MASTER, VMSMAIL_PROFILE, layered product config files (eg. UCX), ...fA better make a copy of these files on every node before the rebooti   --  < Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111-2651; Network and OpenVMS system manager  Fax.    +43 1 81111-888 < FBFV/Information Services           E-mail  eplan@kapsch.netF <<< KAPSCH AG  Wagenseilgasse 1     PSImail PSI%(0232)281001141::EPLANH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"N "VMS is today what Microsoft wants Windows NT V8.0 to be!" Compaq, 22-Sep-1998   ------------------------------  # Date: Wed, 05 Jul 2000 15:33:09 GMTr+ From: Paul Anderson <panderson@genicom.com>  Subject: Re: LA36 Escape codesC Message-ID: <panderson-FD4E76.11332605072000@news.bellatlantic.net>h  D In article <8jpng7$7p5$1@nnrp1.deja.com>, mapryan@my-deja.com wrote:  C > Anyone got a copy of the LA36 escape codes ? We've recently takenaH > delivery of one of these, and it doesn't come with a user manual, just > an installation guide !(  D Is this an old DIGITAL LA36 or the new Compaq LA36 (same name, very  different printers)?   Paul   -- e"    Paul Anderson, DCPS Engineering"    GENICOM Corporation, Gardner MA   ------------------------------   Date: 5 Jul 2000 13:06:54 GMTm+ From: "labadie" <gerard.labadie@compaq.com>s Subject: Re: LAVC Cluster C Message-ID: <01bfe681$e071fe80$521ebc10@w-glabadie.frq.cpqcorp.net>o  B You have in sys$examples a tool (lavc$stop_bus and lavc$start_bus)@ that lets you stop and start the Lavc protocol on any interface.   Grard       : Barry Streets <berrys2552@home.com> a crit dans l'article4 <Oga85.25673$i5.254000@news1.frmt1.sfba.home.com>...J > I have cluster made up of 5 4000/90 ws's, where one is the boot node and the H > others are satellites. I'm looking to reduce disk i/o traffic over theI > ethernet, and to improve performance. (all the nodes have disks mounteduG > cluster wide to share information, and one node is used for serving a  backup > tape drives) > J > If I add the Turbo-channel FDDI Interfaces, will the cluster traffic run" > over it instead of the ethernet? > ! > The systems are running VMS 7.1  >  > Barry Streets  > Echo Management Groupt >  >  >    ------------------------------  $ Date: Wed, 5 Jul 2000 14:47:55 +0100* From: "Richard Brodie" <R.Brodie@rl.ac.uk> Subject: Re: LAVC Cluster., Message-ID: <8jveac$190q@newton.cc.rl.ac.uk>  6 "labadie" <gerard.labadie@compaq.com> wrote in message= news:01bfe681$e071fe80$521ebc10@w-glabadie.frq.cpqcorp.net...f  D > You have in sys$examples a tool (lavc$stop_bus and lavc$start_bus)B > that lets you stop and start the Lavc protocol on any interface.  I At the risk of repeating myself, these should be a method of last resort.iA VMS does a good job at determining the best path dynamically, andlF failing over if necessary: cf the recent post where the LAVC preferred; a 10Mbit Ethernet interface to a misconfigured 100Mbit one.   E As, I understand it, VMS always prefers a FDDI to Ethernet if allowed ; to use full size datagrams. Why disable your fallback path?c   ------------------------------  # Date: Wed, 05 Jul 2000 16:52:33 GMTy1 From: "Mark D. Jilson" <jilly@clarityconnect.com>m Subject: Re: LAVC Clustero2 Message-ID: <396367B4.3A16277E@clarityconnect.com>  ? In my experience a lot of folks equate *any* SCS packets on the ? non-preferred circuit as evidence that VMS is using that as theaE preferred path.  They seem to think that if VMS is using another pathoF then there should be no SCS packets on any other path.  Of course this? isn't true as VMS will still be sending some SCS packets on thenA non-preferred paths for maintenance reasons.  Others have alreadyn; discussed how the chosen packet size will also effect this.    Richard Brodie wrote:e > 8 > "labadie" <gerard.labadie@compaq.com> wrote in message? > news:01bfe681$e071fe80$521ebc10@w-glabadie.frq.cpqcorp.net...n > F > > You have in sys$examples a tool (lavc$stop_bus and lavc$start_bus)D > > that lets you stop and start the Lavc protocol on any interface. > K > At the risk of repeating myself, these should be a method of last resort.uC > VMS does a good job at determining the best path dynamically, andeH > failing over if necessary: cf the recent post where the LAVC preferred= > a 10Mbit Ethernet interface to a misconfigured 100Mbit one.W > G > As, I understand it, VMS always prefers a FDDI to Ethernet if allowedy= > to use full size datagrams. Why disable your fallback path?e   -- gD Jilly	- Working from Home in the Chemung River Valley - Lockwood, NY0 	- jilly@clarityconnect.com			- Brett Bodine fan. 	- Mark.Jilson@Compaq.com			- since 1975 or so, 	- http://www.jilly.baka.com               -   ------------------------------   Date: 5 Jul 2000 08:27:44 GMTe+ From: "labadie" <gerard.labadie@compaq.com>w  Subject: Re: Locate the hotfilesC Message-ID: <01bfe65a$e135ae60$521ebc10@w-glabadie.frq.cpqcorp.net>x   You have a good free tool on   http://www.chez.com/jfp/procio  . It does basically the following, for a process   ana/syso set pro/id=4pid. sh proc/channelb+ then for each file (not .exe or mailbox...)e for 'window2= note the values of wcb$l_reads and wcb$l_writes for this fileu  6 prints the statistics, sorted by reads or writes, e.g.    $ perl procio.pl 20200155 total 
 Reads Writes k7 67802 0 $1$DKA100:[BASES_RDB.MABASE]MABASE_AREA02.RDA; n7 37176 0 $1$DKA100:[BASES_RDB.MABASE]MABASE_AREA01.RDA; n+ 0 1423 $1$DKA100:[USERS.JFP]SORTWORK1.TMP; t/ 1370 0 $1$DKA100:[BASES_RDB.MABASE]MABASE.RDA; ,+ 0 1252 $1$DKA100:[USERS.JFP]SORTWORK0.TMP; i; 455 0 $1$DKA100:[BASES_RDB.MABASE]MABASE_AREA_DEFAULT.RDA; 0: 35 0 $1$DKA100:[BASES_RDB.MABASE]MABASE_AREA_DEFAULT.SNP; - 34 0 $1$DKA100:[BASES_RDB.MABASE]MABASE.SNP; i- 24 7 $1$DKA100:[BASES_RDB.MABASE]MABASE.RDB; o. 5 0 $1$DKA0:[SYSCOMMON.SYSEXE]RIGHTSLIST.DAT; 6 0 5 $1$DKA100:[USERS.RDM$RUJ]MABASE$00014612AA9E.RUJ;  $           / twong@livingstonintl.com a crit dans l'articlek  <8jdaj6$i4g$1@nnrp1.deja.com>...G > One of the disk has much higher i/o rate in comparing with others.  It@ > desire to enable some global buffer counts on those hot files. > ; > Is there any way to locate the files with high i/o rate ?. >  > Thanks for inputs in advance > 
 > Terence. >  > ( > Sent via Deja.com http://www.deja.com/ > Before you buy.u >    ------------------------------  % Date: Wed, 05 Jul 2000 10:13:04 +0100m- From: "POWERS, John" <John.POWERS@sema.co.uk>,  Subject: RE: Locate the hotfiles; Message-ID: <D30A62ABC710D211AEE100A0C9D615EE3163F9@REAES2>o  ; Is this correct? I tried this address and got the message..B  
 erreur!!!! 404d Document non trouv=E9i  L Even with my limited linguistic skills can translate that (everything except the 404)   - John   > -----Original Message-----2 > From: labadie [mailto:gerard.labadie@COMPAQ.COM] > Sent: 05 July 2000 09:28 > To: Info-VAX@Mvb.Saic.Comr" > Subject: Re: Locate the hotfiles >=20 >=20 > You have a good free tool on >=20  > http://www.chez.com/jfp/procio >=200 > It does basically the following, for a process >=20	 > ana/sysn > set pro/id=3D4pid  > sh proc/channele- > then for each file (not .exe or mailbox...)l
 > for 'window ? > note the values of wcb$l_reads and wcb$l_writes for this file/ >=208 > prints the statistics, sorted by reads or writes, e.g. >=20$ > $ perl procio.pl 20200155 total=20 > Reads Writes=20n; > 67802 0 $1$DKA100:[BASES_RDB.MABASE]MABASE_AREA02.RDA;=20t; > 37176 0 $1$DKA100:[BASES_RDB.MABASE]MABASE_AREA01.RDA;=20m/ > 0 1423 $1$DKA100:[USERS.JFP]SORTWORK1.TMP;=20o3 > 1370 0 $1$DKA100:[BASES_RDB.MABASE]MABASE.RDA;=20a/ > 0 1252 $1$DKA100:[USERS.JFP]SORTWORK0.TMP;=20i? > 455 0 $1$DKA100:[BASES_RDB.MABASE]MABASE_AREA_DEFAULT.RDA;=20f> > 35 0 $1$DKA100:[BASES_RDB.MABASE]MABASE_AREA_DEFAULT.SNP;=201 > 34 0 $1$DKA100:[BASES_RDB.MABASE]MABASE.SNP;=20 1 > 24 7 $1$DKA100:[BASES_RDB.MABASE]MABASE.RDB;=20o2 > 5 0 $1$DKA0:[SYSCOMMON.SYSEXE]RIGHTSLIST.DAT;=20: > 0 5 $1$DKA100:[USERS.RDM$RUJ]MABASE$00014612AA9E.RUJ;=20 > $=20 >=20 >=20 >=20 >=20 >=203 > twong@livingstonintl.com a =E9crit dans l'articled" > <8jdaj6$i4g$1@nnrp1.deja.com>...A > > One of the disk has much higher i/o rate in comparing with=20  > others.  IB > > desire to enable some global buffer counts on those hot files. > >=20= > > Is there any way to locate the files with high i/o rate ?l > >=20  > > Thanks for inputs in advance > >=20 > > Terence. > >=20 > >=20* > > Sent via Deja.com http://www.deja.com/ > > Before you buy.  > >=20 >=20    K ___________________________________________________________________________aD This email is confidential and intended solely for the use of the=20J individual to whom it is addressed. Any views or opinions presented are=20G solely those of the author and do not necessarily represent those of=20F Sema Group.=20L If you are not the intended recipient, be advised that you have received th= isK email in error and that any use, dissemination, forwarding, printing, or=20e- copying of this email is strictly prohibited.h  E If you have received this email in error please notify the Sema Groupt. Helpdesk by telephone on +44 (0) 121 627 5600.K ___________________________________________________________________________a   ------------------------------  % Date: Wed, 05 Jul 2000 06:22:13 -0400)2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>  Subject: RE: Locate the hotfiles7 Message-ID: <200007050622_MC2-AB23-D8A9@compuserve.com>   J         404 is "File not found".  It has also become slang for "clueless"= ;r "he's 404 on that".e  & Message text written by "POWERS, John"< >Is this correct? I tried this address and got the message..  
 erreur!!!! 404e Document non trouv=E9l  E Even with my limited linguistic skills can translate that (everything  except	 the 404)<t   ------------------------------   Date: 5 Jul 2000 12:11:44 GMT-+ From: "labadie" <gerard.labadie@compaq.com>r  Subject: Re: Locate the hotfilesC Message-ID: <01bfe67a$2b837fa0$521ebc10@w-glabadie.frq.cpqcorp.net>    Try this url# http://www.chez.com/jfp/procio.htmlS or http://www.chez.com/jfp 
 and click on a procio at the bottom left -   Sorry for the inconvenienceu   Grard      < POWERS, John <John.POWERS@sema.co.uk> a crit dans l'article2 <D30A62ABC710D211AEE100A0C9D615EE3163F9@REAES2>...; Is this correct? I tried this address and got the message..i  
 erreur!!!! 404L Document non trouvc  E Even with my limited linguistic skills can translate that (everythingo except the 404)   - John   > -----Original Message-----2 > From: labadie [mailto:gerard.labadie@COMPAQ.COM] > Sent: 05 July 2000 09:28 > To: Info-VAX@Mvb.Saic.Comt" > Subject: Re: Locate the hotfiles >  >  > You have a good free tool on >   > http://www.chez.com/jfp/procio > 0 > It does basically the following, for a process > 	 > ana/sysl > set pro/id=4pid  > sh proc/channel[- > then for each file (not .exe or mailbox...)l
 > for 'window-? > note the values of wcb$l_reads and wcb$l_writes for this file1 > 8 > prints the statistics, sorted by reads or writes, e.g. > " > $ perl procio.pl 20200155 total  > Reads Writes F9 > 67802 0 $1$DKA100:[BASES_RDB.MABASE]MABASE_AREA02.RDA;  9 > 37176 0 $1$DKA100:[BASES_RDB.MABASE]MABASE_AREA01.RDA; F- > 0 1423 $1$DKA100:[USERS.JFP]SORTWORK1.TMP; s1 > 1370 0 $1$DKA100:[BASES_RDB.MABASE]MABASE.RDA; U- > 0 1252 $1$DKA100:[USERS.JFP]SORTWORK0.TMP; a= > 455 0 $1$DKA100:[BASES_RDB.MABASE]MABASE_AREA_DEFAULT.RDA; a< > 35 0 $1$DKA100:[BASES_RDB.MABASE]MABASE_AREA_DEFAULT.SNP; / > 34 0 $1$DKA100:[BASES_RDB.MABASE]MABASE.SNP; r/ > 24 7 $1$DKA100:[BASES_RDB.MABASE]MABASE.RDB; c0 > 5 0 $1$DKA0:[SYSCOMMON.SYSEXE]RIGHTSLIST.DAT; 8 > 0 5 $1$DKA100:[USERS.RDM$RUJ]MABASE$00014612AA9E.RUJ;  > $  >  >  >  >  > 1 > twong@livingstonintl.com a crit dans l'article " > <8jdaj6$i4g$1@nnrp1.deja.com>...? > > One of the disk has much higher i/o rate in comparing with n > others.  IB > > desire to enable some global buffer counts on those hot files. > > = > > Is there any way to locate the files with high i/o rate ?G > >   > > Thanks for inputs in advance > >  > > Terence. > >  > > * > > Sent via Deja.com http://www.deja.com/ > > Before you buy.. > >  >     K ___________________________________________________________________________mB This email is confidential and intended solely for the use of the H individual to whom it is addressed. Any views or opinions presented are E solely those of the author and do not necessarily represent those of s Sema Group. H If you are not the intended recipient, be advised that you have received thisI email in error and that any use, dissemination, forwarding, printing, or g- copying of this email is strictly prohibited.=  E If you have received this email in error please notify the Sema GroupE. Helpdesk by telephone on +44 (0) 121 627 5600.K ___________________________________________________________________________=  
 ----------   ------------------------------  / Date: Wed, 05 Jul 2000 08:47:02 +0200 (MET DST) & From: Rudolf Wingert <win@fom.fgan.de>9 Subject: Missing pseudocolor visual (Elsa Gloria Synergy)R/ Message-ID: <200007050553.HAA24180@fom.fgan.de>b   Hellod  H we do have a few AlphaStation DS20 with the Elsa Gloria Synergy graphic-H adapter. The X11 server under OpenVMS 7.1-2 AXP (DECwindows Motif 1.2-4)E starts the display without a pseudocolor visual. Our application withpG thousand of lines code, uses the pseudocolor visual, to display images. C So we can't use them on the DS AlphaStation. Is there any solution?wG Am I able to config an additional pseudocolor visual with the procedure  DECW$PRIVATE_SERVER_SETUP.COM?   TIA und regards Rudolf Wingert   ------------------------------  % Date: Wed, 05 Jul 2000 16:05:46 +0200l. From: Nichele Claudio <claudio.nichele@nbb.be>, Subject: Multilingual System Dictionary(ies)> Message-ID: <000301bfe68a$1ed89910$0457c80a@pc2860.lanprd.nbb>  
 Hello all,F I would like to use on all our production systems the french and dutch? System Dictionaries in conjunction with the standard dictionarym3 available SYS$LIBRARY:VMS$PASSWORD_DICTIONARY.DATA.iB But i am unaware of how to do it and the official documentation is poor in informations.fF Are System Dictionaries for other languages than english available? If) yes, where can i found - download - them? 0 And how to intall these new System Dictionaries?C The best would be to maintain System Dictionaries in separate .DATAuB files, and, for example, using them with a search list in a system logical. Is this possible?D The worst approach is to merge all dictionaries in an unique and big  .DATA file with all words mixed.' We are using OpenVMS V7.1 and V7.1-1H2.  Thanks for your help.-    ;       --------- [ All views expressed are my own ] --------k Claudio Nichelee4 OpenVMS System Programmer - National Bank of Belgium Tel: +32 (0)2 221.28.11  Email: claudio.nichele@nbb.be8*                               ------------                    !"DISCLAIMER: The content of this e-mail message does not constitute a commitment of the National Bank of Belgium (NBB) except where provided for in a written agreement between you and the NBB or where confirmed with a written form approved according to the internal regulations of the NBB.A Besides, the statements and opinions expressed in this e-mail message are those of the author of the message and do not necessarily represent those of the NBB.-BThe e-mail message contains proprietary information intended for the intended recipient only. If an addressing or transmission error has misdirected this e-mail, please notify the author. If you are not the intended recipient you must not use, disclose, distribute, copy, print or rely on any part of this e-mail message."   ------------------------------   Date: 5 Jul 2000 11:42:18 -0500S9 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)o0 Subject: Re: Multilingual System Dictionary(ies)+ Message-ID: <mS5g1F6v783O@eisner.decus.org>l  o In article <000301bfe68a$1ed89910$0457c80a@pc2860.lanprd.nbb>, Nichele Claudio <claudio.nichele@nbb.be> writes:t  H > I would like to use on all our production systems the french and dutchA > System Dictionaries in conjunction with the standard dictionaryt5 > available SYS$LIBRARY:VMS$PASSWORD_DICTIONARY.DATA.RD > But i am unaware of how to do it and the official documentation is > poor in informations.e  7 Fiddling with VMS$PASSWORD_DICTIONARY is not supported.b1 Fiddling with VMS$PASSWORD_DICTIONARY is trivial.e  H > Are System Dictionaries for other languages than english available? If+ > yes, where can i found - download - them?   D If Compaq supports any (and I have never heard of it) it would be inF their regional offices (Belgium in your case).  The central developersE just do English (and provide the hooks so the regional offices can doN replacements).  2 > And how to intall these new System Dictionaries?E > The best would be to maintain System Dictionaries in separate .DATAgD > files, and, for example, using them with a search list in a system > logical. Is this possible?   No, the code does not do that.  F > The worst approach is to merge all dictionaries in an unique and big" > .DATA file with all words mixed.  C So fiddling with VMS$PASSWORD_DICTIONARY is not even a goal in youruD case.  You can write a VMS Password Policy module following pointersJ starting in 7.3.3.3 of AA-Q2HLD-TE, The VMS V7.2 Guide to System Security.   ------------------------------  # Date: Wed, 05 Jul 2000 10:18:06 GMTd From: pavichal@my-deja.com- Subject: Netscape Fasttrack for OpenVMS 7.2-1 ) Message-ID: <8jv20m$upc$1@nnrp1.deja.com>s  ; I have just installed Netscape FastTrack for OpenVMS 7.2-1.1H I have been trying to get it to run DCL CGIs but dont seem to be able toE get it to do anything. All it does is shows me the script that I wantrA it to run. I have tried a few simple examples that I found on the5) Digital website, but they will not work :2  , $ write sys$output "Content-type: text/html" $ write sys$output ""oA $ write sys$output "                    HELLO                   "e $ write sys$output ""g  F Have I missed something out in the setup or is there something missingF in the script. I have setup Fasttrack to assume all files in a cgi are directory are for cgi.   Cheers.n   Piyush.p    & Sent via Deja.com http://www.deja.com/ Before you buy.f   ------------------------------  # Date: Wed, 05 Jul 2000 10:58:42 GMT ' From: Colin Blake <colin@theblakes.com> 1 Subject: Re: Netscape Fasttrack for OpenVMS 7.2-1 - Message-ID: <396314E1.1F530648@theblakes.com>e   pavichal@my-deja.com wrote:r  H > Have I missed something out in the setup or is there something missingH > in the script. I have setup Fasttrack to assume all files in a cgi are > directory are for cgi.  C That should work. I also run with "CGI File Type" enabled, but thata shouldn't be necessary.   # You did "apply" the changes, right?h   ------------------------------  # Date: Wed, 05 Jul 2000 12:56:03 GMTr From: pavichal@my-deja.com1 Subject: Re: Netscape Fasttrack for OpenVMS 7.2-1g) Message-ID: <8jvb8r$54q$1@nnrp1.deja.com>c  - In article <396314E1.1F530648@theblakes.com>,e*   Colin Blake <colin@theblakes.com> wrote: > pavichal@my-deja.com wrote:g >dB > > Have I missed something out in the setup or is there something missingiF > > in the script. I have setup Fasttrack to assume all files in a cgi are- > > directory are for cgi. > E > That should work. I also run with "CGI File Type" enabled, but thatr > shouldn't be necessary.R >L% > You did "apply" the changes, right?6 >E  B I am actually running with CGI Directory so that everything in theG directory is assumed to be for CGI. I have applied thae changes, but ith8 still just prints out the script rather then running it. >s    & Sent via Deja.com http://www.deja.com/ Before you buy.a   ------------------------------   Date: 5 Jul 2000 16:08:17 GMTd2 From: mathog@seqaxp.bio.caltech.edu (David Mathog); Subject: Re: OpenVMS loses big, was:  RE: Compaq advertisesh+ Message-ID: <8jvmhh$b7@gap.cco.caltech.edu>   Y In article <39609D29.2CEDAEC2@bbc.co.uk>, Tim Llewellyn <tim.llewellyn@bbc.co.uk> writes:s >b >What I am saying isJ >"sure, I knew that, to get performance on VMS design your application for >VMS,especially, don'tG >use zillions of little files".  Of course, if you can do configurationf# >management properly and understandmO >comupting rather than just a limited aspect of one particlar operating system,i >it doesn't seem tonH >like that much hard work to make an application work efficiently with a >consistent, userfriendlyEQ >(non-unix) user interface on a range of platforms including VMS and unix. Oraclea >seem to be  >a good example.  " BZZZZZT.  Wrong answer, try again.  G For a UNIT text file operation, going RAMDISK to RAMDISK on OpenVMS, ornI file cache to file cache on Linux, on identical DS10s, the result is thatrH the Linux system is 2.5-6.5 times faster.  This is for an operation likeH "read text record from input, write text record to output" - pure IO. ItJ doesn't matter if you do this in 1 file or in 1000 you're already startingK out with the Unix systems "lighter" text file handling mechanisms 3X fasterlI than those on OpenVMS.  And it goes downhill from there, rapidly, because.K the lack of file effective file caching on OpenVMS throws in another factor H of 10 advantage to Linux.   The only time I can get similar performance I from my VMS box is when the IO is minimal (a CPU bound program) or the IOhJ is done differently, via memory mapping or some other mechanism to bypass  RMS.).   >2I >David, how many of your "applications" were written by grad students andE >postsdocs?7  G Oh, most of them.   That's the reality I've got to live with.  However,SK while the quality of the code may vary, I cannot lay the blame there. Their F programming approaches are reasonable and work well on other OS's. TheJ fault is with OpenVMS, which simply has atrocious IO performance when usedJ on small systems with reasonably small data files, even AFTER you've gone G to RAMdisks.  Take a look at the benchmarks I posted a while back - theSJ only way to speed them up on OpenVMS is to avoid fprintf() and instead useG routines which bypass both the C RTL and RMS and go straight to the QIOpI level.  That is unacceptable - virtually all software now is written in ChI (or C++, which I have not tested but assume has the same problems) and iteG is completely unreasonable that the OpenVMS user should have to rewrite B reasonably written code so completely in order to achieve adequate performance on this OS.    David Mathog mathog@seqaxp.bio.caltech.edu-? Manager, sequence analysis facility, biology division, Caltech WJ **************************************************************************J *                                RIP VMS                                 *J **************************************************************************   ------------------------------  # Date: Wed, 05 Jul 2000 15:18:34 GMT . From: Michael Austin <maustin@nc.prestige.net>; Subject: Re: OpenVMS loses big, was:  RE: Compaq advertisese/ Message-ID: <39635F6C.2E4D3BEB@nc.prestige.net>n  , This is a multi-part message in MIME format.& --------------8DED81F499063B55FE257E1F* Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bitl       David Mathog wrote:e  [ > In article <39609D29.2CEDAEC2@bbc.co.uk>, Tim Llewellyn <tim.llewellyn@bbc.co.uk> writes:o > >l > >What I am saying isL > >"sure, I knew that, to get performance on VMS design your application for > >VMS,especially, don'tI > >use zillions of little files".  Of course, if you can do configuratione% > >management properly and understandSQ > >comupting rather than just a limited aspect of one particlar operating system,A > >it doesn't seem to3J > >like that much hard work to make an application work efficiently with a > >consistent, userfriendly$S > >(non-unix) user interface on a range of platforms including VMS and unix. OracleE
 > >seem to be2 > >a good example. >A$ > BZZZZZT.  Wrong answer, try again. >0I > For a UNIT text file operation, going RAMDISK to RAMDISK on OpenVMS, orEK > file cache to file cache on Linux, on identical DS10s, the result is that=J > the Linux system is 2.5-6.5 times faster.  This is for an operation likeJ > "read text record from input, write text record to output" - pure IO. ItL > doesn't matter if you do this in 1 file or in 1000 you're already startingM > out with the Unix systems "lighter" text file handling mechanisms 3X fasterfK > than those on OpenVMS.  And it goes downhill from there, rapidly, becauselM > the lack of file effective file caching on OpenVMS throws in another factor I > of 10 advantage to Linux.   The only time I can get similar performanceeK > from my VMS box is when the IO is minimal (a CPU bound program) or the IO_K > is done differently, via memory mapping or some other mechanism to bypasse > RMS.)e >e  [ Of course UNIX will always be faster when doing pure I/O... you get much better performancet\ on a system with ZERO error checking (used to make sure you are not overwriting something onA the disk... which is still possible in some unix environments...)s   my $.02 worth.   >a > > K > >David, how many of your "applications" were written by grad students andr
 > >postsdocs?d > I > Oh, most of them.   That's the reality I've got to live with.  However,eM > while the quality of the code may vary, I cannot lay the blame there. Their_H > programming approaches are reasonable and work well on other OS's. TheL > fault is with OpenVMS, which simply has atrocious IO performance when usedK > on small systems with reasonably small data files, even AFTER you've gone2I > to RAMdisks.  Take a look at the benchmarks I posted a while back - thesL > only way to speed them up on OpenVMS is to avoid fprintf() and instead useI > routines which bypass both the C RTL and RMS and go straight to the QIOeK > level.  That is unacceptable - virtually all software now is written in CdK > (or C++, which I have not tested but assume has the same problems) and it-I > is completely unreasonable that the OpenVMS user should have to rewriteeD > reasonably written code so completely in order to achieve adequate > performance on this OS.5 >b > David Mathog > mathog@seqaxp.bio.caltech.edu @ > Manager, sequence analysis facility, biology division, Caltech > **7   ------------------------------   Date: 5 Jul 2000 13:42:32 -0500u9 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)c; Subject: Re: OpenVMS loses big, was:  RE: Compaq advertisesE+ Message-ID: <LoKsvv2fycO9@eisner.decus.org>s  ` In article <8jvmhh$b7@gap.cco.caltech.edu>, mathog@seqaxp.bio.caltech.edu (David Mathog) writes:[ > In article <39609D29.2CEDAEC2@bbc.co.uk>, Tim Llewellyn <tim.llewellyn@bbc.co.uk> writes:g >> >>What I am saying isiK >>"sure, I knew that, to get performance on VMS design your application forb >>VMS,especially, don't:H >>use zillions of little files".  Of course, if you can do configuration$ >>management properly and understandP >>comupting rather than just a limited aspect of one particlar operating system, >>it doesn't seem toI >>like that much hard work to make an application work efficiently with a  >>consistent, userfriendlyR >>(non-unix) user interface on a range of platforms including VMS and unix. Oracle >>seem to be >>a good example.a > $ > BZZZZZT.  Wrong answer, try again.  / Polite discourse would be much better received.1  J >>David, how many of your "applications" were written by grad students and >>postsdocs? > I > Oh, most of them.   That's the reality I've got to live with.  However,RM > while the quality of the code may vary, I cannot lay the blame there. Their.H > programming approaches are reasonable and work well on other OS's. TheL > fault is with OpenVMS, which simply has atrocious IO performance when usedL > on small systems with reasonably small data files, even AFTER you've gone I > to RAMdisks.  Take a look at the benchmarks I posted a while back - the L > only way to speed them up on OpenVMS is to avoid fprintf() and instead useI > routines which bypass both the C RTL and RMS and go straight to the QIO K > level.  That is unacceptable - virtually all software now is written in C.K > (or C++, which I have not tested but assume has the same problems) and it I > is completely unreasonable that the OpenVMS user should have to rewritevD > reasonably written code so completely in order to achieve adequate > performance on this OS.   C You do seem to claim that your complaint is with VMS in general andnC only incidentally does it slip out that your concerns are only withsC C.  You blame RMS and they reveal that you are actually calling RMS  through the C Runtime library.  ? Try some tests in Ada, Pascal or Fortran, or else make it quitei> clear in your strident posts that yours is a C-only viewpoint.   ------------------------------  $ Date: Wed, 5 Jul 2000 13:47:26 -0400' From: "Bill Todd" <billtodd@foo.mv.com>h; Subject: Re: OpenVMS loses big, was:  RE: Compaq advertisesE' Message-ID: <8jvs8j$g4$1@pyrite.mv.net>7  D Larry Kilgallen <Kilgallen@eisner.decus.org.nospam> wrote in message% news:LoKsvv2fycO9@eisner.decus.org...-   ...-  E > You do seem to claim that your complaint is with VMS in general anduE > only incidentally does it slip out that your concerns are only withsE > C.  You blame RMS and they reveal that you are actually calling RMS   > through the C Runtime library. >oA > Try some tests in Ada, Pascal or Fortran, or else make it quitee@ > clear in your strident posts that yours is a C-only viewpoint.  I This observation may be reasonable with respect to wording of the post tooJ which it responded, but should not be taken as any indication that defaultK performance using RMS does not have similar problems - since it does indeedoD have very similar problems.  They can be alleviated significantly byJ explicitly taking advantage of optional RMS features, but the result stillI often falls considerably short of the performance one can achieve on UnixeI just using default settings (and without sacrifice in integrity, since in0K both environments judicious use of $FLUSH or fsync is required at any pointsJ where a guarantee that written data is actually on the disk is important).   - bill   ------------------------------  + Date: Wed, 05 Jul 2000 09:37:59 +0000 (GMT)A) From: Yosef Branse <JODY@lib.haifa.ac.il>t% Subject: Problem with VFYMAIL utilityn. Message-ID: <000705093759.375@lib.haifa.ac.il>   Hello, h  ( I am running VMS 7.2-1 on an Alpha 4100.  F I recently tried to use the program VFYMAIL, available on the Freeware5 distribution, in order to fix a corrupted mail file. s  J At first, I used the precompiled executable, and followed the instructionsF included in the file FREEWARE_DEMO.TXT included with the distribution:  E >   1. COPY the appropriate file (VFYMAIL.EXE_VAX or VFYMAIL.EXE_AXP)u >      to VFYMAIL.EXE. >17 >   2. Edit the following line in the VFYMAIL.CLD file:i >  >         image       "VFYMAIL"  >uO >      to point to the directory containing the VFYMAIL.EXE file.  For example,  >      "SYS$DISK:[]VFYMAIL". >e! >   3. Type "SET COMMAND VFYMAIL"  >  >   4. Type "VFYMAIL /STAT"   0 However, I received an error message as follows:  N ******************************************************************************@ %VFYMAIL-I-PROMAIDIR, processing mail directory USER:[JODY.MAIL]> %VFYMAIL-I-BYPSEQMAI, bypassing sequential mailfile KEYDEF.MAI2 %VFYMAIL-I-PROMAIFIL, processing mailfile MAIL.MAI# %BAS-F-TOOFEWARG, Too few argumentsm; -BAS-I-FROFUN, In external function BUILD_INTERNAL_NAMELIST   -BAS-I-FROMOD, In module VFYMAIL/ %TRACE-F-TRACEBACK, symbolic stack dump followsnJ   image    module    routine             line      rel PC           abs PCO                                             0 FFFFFFFF8045F5EC FFFFFFFF8045F5ECnO  DEC$BASRTL                                 0 000000000000DB2C 000000007C041B2Cr= ----- above condition handler called with exception 001A830C:-# %BAS-F-TOOFEWARG, Too few arguments1 ----- end of exception messageO                                             0 FFFFFFFF8917C67C FFFFFFFF8917C67C(/  VFYMAIL  VFYMAIL_SUBS  MAILFILE_FOLDER_ROUTINElO                                             1 0000000000000D80 00000000000233C0oO  MAILSHR                                    0 00000000000414C4 000000007B6AB4C4 O  MAILSHR                                    0 000000000003F9E4 000000007B6A99E4aO  MAILSHR                                    0 0000000000031374 000000007B69B374mO  MAILSHR                                    0 0000000000031694 000000007B69B694s/  VFYMAIL  VFYMAIL_TREE  BUILD_INTERNAL_NAMELISTt  O                                           934 0000000000000594 0000000000021D54cO                                             0 FFFFFFFF804643EC FFFFFFFF804643ECdO  VFYMAIL  VFYMAIL  VFYMAIL               1991 00000000000010CC 00000000000210CCwO                                             0 FFFFFFFF892E53D4 FFFFFFFF892E53D4sN ******************************************************************************  E Next, I tried to build the program from the source code in the [.SRC]tD subdirectory, using the instructions in the file BUILD_SOFTWARE.TXT:   >    $ BASIC  vfymail_main >    $ BASIC  vfymail_misc >    $ BASIC  vfymail_subs >    $ BASIC  vfymail_tree >    $ LINK  vfymail/OPTIONr  K The compilations worked OK except for that of vfymail_subs, which gave thiso result:.  N ******************************************************************************         BASIC  vfymail_subso  "         xab::xab$b_cod = xab$c_pro ^i> %BASIC-E-EXPDECREQ, explicit declaration of XAB$C_PRO required@ at line number 334 in file UTIL$:[VFYMAIL.SRC]VFYMAIL_SUBS.BAS;1  %         xab::xab$b_bln = xab$c_prolenj ^:A %BASIC-E-EXPDECREQ, explicit declaration of XAB$C_PROLEN requiredn@ at line number 335 in file UTIL$:[VFYMAIL.SRC]VFYMAIL_SUBS.BAS;1  K         xab::xab$w_pro = xab::xab$w_pro AND NOT ((2^xab_v_own)*xab$m_nodel)a	 ........^ . %BASIC-E-ERRRECCOM, erroneous RECORD component@ at line number 341 in file UTIL$:[VFYMAIL.SRC]VFYMAIL_SUBS.BAS;1  K         xab::xab$w_pro = xab::xab$w_pro AND NOT ((2^xab_v_own)*xab$m_nodel)$ .........................^. %BASIC-E-ERRRECCOM, erroneous RECORD component@ at line number 341 in file UTIL$:[VFYMAIL.SRC]VFYMAIL_SUBS.BAS;1  K         xab::xab$w_pro = xab::xab$w_pro AND NOT ((2^xab_v_own)*xab$m_nodel)g ^n@ %BASIC-E-EXPDECREQ, explicit declaration of XAB$M_NODEL required@ at line number 341 in file UTIL$:[VFYMAIL.SRC]VFYMAIL_SUBS.BAS;1  K         xab::xab$w_pro = xab::xab$w_pro AND NOT ((2^xab_v_own)*xab$m_nodel)-- ............................................^ = %BASIC-E-LOGOPENON, logical operation on non-integer quantityr@ at line number 341 in file UTIL$:[VFYMAIL.SRC]VFYMAIL_SUBS.BAS;1  P %BASIC-E-ENDNOOBJ, UTIL$:[VFYMAIL.SRC]VFYMAIL_SUBS.BAS;1 completed with 6 diagno stics - object deletedN ******************************************************************************  M Although I am wary of trying to fix a program that I didn't write, I took theb2 liberty of adding these lines to VFYMAIL_SUBS.BAS:       DECLARE LONG xab$c_pro, &-                   xab$c_prolen, &                  xab$m_nodel  < This time, the compilation did not return errors of the sort  F      %BASIC-E-EXPDECREQ, explicit declaration of XAB$M_NODEL required,  K but the other errors I'd received when compiling the unmodified source werer still there.  J Searching the archives of this list in DEJA-NEWS, I saw that the topic wasL discussed here a few months ago, and apparently the problem arose only after4 upgrading to VMS 7.2, but no solution was provided.   9 In his letter of 24 December 1999, Stephen Hoffman wrote:b  O >  I'd suggest a look at the sources and a rebuild, as it looks like something tN >  is broken -- though I'd not expect anything that was using only documented O >  interfaces to encounter this sort of problem...  So this problem is a resulttM >  of either a busted RTL on OpenVMS, or something in the code that is using /. >  an undocumented or unsupported interface... >-J >  I'll queue a note to myself look at this for the next Freeware release K >  cycle.  If somebody else figures this out and fixes this before the nextvK >  release cycle, please let me know and I will update the Freeware copy...9  H Does the above information help clarify what the problem is? Any idea asJ to what needs to be fixed in VFYMAIL_SUBS.BAS? I'd be happy to try out any suggestion.i   Thanks for your assistance,   L ****************************************************************************L * Yosef (Jody) Branse       University of Haifa Library                    *L * Systems Librarian         Mt. Carmel, Haifa 31905, Israel                *L * Webmaster                 Tel.: 972 4-8240288  / FAX:  972 4-8257753     *L *                           WWW Server: http://www-lib.haifa.ac.il         *L *                                                                          *L *            Electronic mail:   JODY@LIB.HAIFA.AC.IL  (VMS Mail)           *L *                               JODY@UNIV.HAIFA.AC.IL (University of Haifa *L *                                                      Exchange Server)    *L *                                                                          *L ****************************************************************************   ------------------------------  / Date: Wed, 05 Jul 2000 08:54:30 +0200 (MET DST)M& From: Rudolf Wingert <win@fom.fgan.de>7 Subject: Problems with the DIRECTORY command (it loops)n/ Message-ID: <200007050600.IAA24817@fom.fgan.de>g   Hello,  C one user of our institute have an application, which produces a loteA of files with the same name, but different versions. Yesterday her@ reports an infinite loop with the DIRECTORY command. From a fileE TMP.DAT do exists about 15800 versions. The DIRECTORY TMP.DAT commandsC starts to display with the highest version down to a version 11xyz,IC displays a non existing version TMP.DAT;32767 and starts displayinga< again with the highest number and so on and so on ... Also aF DIRECTORY/TOTAL command does never come to an end. The OpenVMS versionE is 7.1-2AXP. Is this a known bug? Is the problem to have lot versionss: of a file or to have lot of files in a directory (>32000)?   TIA and regards Rudolf Wingert   ------------------------------    Date: 05 Jul 2000 11:10:03 +0200G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>t; Subject: Re: Problems with the DIRECTORY command (it loops) H Message-ID: <y4sntpvt78.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  ( Rudolf Wingert <win@fom.fgan.de> writes:  E > one user of our institute have an application, which produces a lot C > of files with the same name, but different versions. Yesterday hecB > reports an infinite loop with the DIRECTORY command. From a fileG > TMP.DAT do exists about 15800 versions. The DIRECTORY TMP.DAT commandoE > starts to display with the highest version down to a version 11xyz,*E > displays a non existing version TMP.DAT;32767 and starts displaying > > again with the highest number and so on and so on ... Also aH > DIRECTORY/TOTAL command does never come to an end. The OpenVMS versionG > is 7.1-2AXP. Is this a known bug? Is the problem to have lot versionsg< > of a file or to have lot of files in a directory (>32000)?  E From a note in an ECO release note, I gathered that the new directorytH reshuffle code (part of the stuff that makes large directories bearable)I uses a file name with version=32767 and an impossible file id as a marker K in case if a crash at an inopportune moment. It is supposed to repair that  G the next time it accesses that directory, but as I read this in an ECO,s sometimes this can go wrong...   	Jan   ------------------------------  % Date: Wed, 05 Jul 2000 13:44:09 -0500l* From: WILLIAM WEBB <WWEBB1@email.usps.gov>' Subject: Re[2]: VMS Printing using PCL6r- Message-ID: <0033000000273734000002L042*@MHS>s  	 =0ABarry,s  E We do a lot of embedding of PCL5 in print jobs from VMS.  We just got A some new printers that support PCL6.  I was curious if you have a C complete command reference for PCL6 and if so where did you get it.   ) ---###--- Begin response by WWW ---###---    Brent:     First, the bad news-  H From  <http://www.hp.com/cposupport/printers/support_doc/bpl04568.html>= :   H NOTE: PCL6 is very much different from PCL5 and previous PCL versions. = One H significant difference is that you cannot use PCL6 to send commands to = ther  printer in the same way as PCL5.     Then the good news:i  6 From  <http://www.hp.com/pressrel/apr96/08apr96b.htm>:  H PCL 6 is completely backward-compatible with today's software applicati= onst4 because PCL 5e is integrated directly into HP PCL 6.  F Not what you wanted in a response, but useful information nonetheless.     WWWebb  ' ---###--- End response by WWW ---###---h   Thanks  
 Brent Wilhelmo    & Sent via Deja.com http://www.deja.com/ Before you buy.a =e   ------------------------------  % Date: Wed, 05 Jul 2000 13:22:18 -0400b- From: JF Mezei <jfmezei.spamnot@videotron.ca>,. Subject: Re: RMS indexed file instatly growing, Message-ID: <39636EC6.BD154BCE@videotron.ca>   Randy Park wrote:o1 > Several things you should understand about RMS:nB > 1.  You will have less bucket splitting, and thus a smaller file@ > size, if your records are added in primary key sequence order.E > 2.  Even after you delete a record, the RFA (Record's File Address)a7 > or the retrieval pointer for the record still exists.nA > 3.  The CONVERT/RECLAIM command will recover space in the file.v  L The way it was explained to me is that if you delete all records in a bucketH before it needs to be split, then the bucket becomes available again andI convert/reclaim is not needed. But if you don't, then unused space in thetT bucket (deleted records) is not made available and you're stuck with a growing file.  L You should note that convert/reclaim is callable from inside a program. Just< ensure the file is closed and call the conv$reclaim routine:  $ $DESCRIPTOR(myfile,"MYBIFILE.DATA"); struct { long number;o 	long data_scanned ; 	long data_reclaimed ; 	long index_reclaimed ;c 	long total_reclaimed ;s
 	} myvector ;w   myvector.number = 4 ) status = conv$reclaim(&myfile,&myvector);   < and voila. This also drops statistics in the myvector thing.   ------------------------------  # Date: Wed, 05 Jul 2000 16:35:06 GMTl1 From: "Mark D. Jilson" <jilly@clarityconnect.com>i4 Subject: Re: RMS tuning breaking programs, any ideas2 Message-ID: <3963639C.996AEC31@clarityconnect.com>  G You are exceeding you PIOPAGES area.  If you have access to the DSNLINKuE knowledge base then searching on PIOPAGES & RMS will give you all theKF information that you will need to properly configure PIOPAGES for your desired RMS values.2   David Mathog wrote:s > H > Can anybody hazard a guess why raising BLOCK and BUFFER together wouldD > cause a program to fail, but raising each alone caused no problem? >  > $ sho rmslO >           MULTI-  |                MULTIBUFFER COUNTS               | NETWORKoN >           BLOCK   | Indexed  Relative            Sequential         |  BLOCKN >           COUNT   |                     Disk   Magtape  Unit Record |  COUNTL > Process   127     |    0         0        0       0         0       |    0L > System     16     |    0         0        0       0         0       |    8 > 2 >           Prolog    Extend Quantity      VCC_DFW/ > Process     0              0                0e/ > System      0              0                0n > $!, > $! the next "fetch" command works properly > $! > $ fetch pir1:a1hu  > $!# > $! the next "fetch" command failst > $! > $ set rms/buffer=255 > $ fetch pir1:a1hua > $!, > $! the next "fetch" command works properly > $! > $ set rms/block=0  > $ fetch pir1:a1huI > G > It looks like setting block AND buffer up to their maximum is causing F > something in this program to fail, but setting either one alone thatG > high has no effect.   "fetch" is part of the GCG package, and it goeslD > N levels deep in libraries, and is commercial.  I've also recentlyE > seen this same block+buffer=BOOM effect when sending mail using EDTa@ > as the editor.  On trying to exit from EDT I was greeted with: > " > Output file could not be created6 > File name:USRDISK:[USERS.MATHOG]MAIL_394E_SEND.TMP;1& > %RMS-F-DME, dynamic memory exhausted > E > Which made no sense to me at all, since at the time there was a lotn' > of free dynamic memory in the system:d > N > Dynamic Memory Usage (bytes):      Total        Free      In Use     LargestN >   Nonpaged Dynamic Memory        2711552      625792     2085760      257536N >   Paged Dynamic Memory           2932736     1199872     1732864     1194720B >   Lock Manager Dynamic Memory    1384448      688128      696320 > ) > The process where this all happend has:  > ; > Maxjobs:         0  Fillm:       100  Bytlm:        50000M; > Maxacctjobs:     0  Shrfillm:      0  Pbytlm:           0P; > Maxdetach:       0  BIOlm:       100  JTquota:       2048e; > Prclm:          10  DIOlm:       100  WSdef:          150-; > Prio:            4  ASTlm:       102  WSquo:          256 ; > Queprio:         0  TQElm:        40  WSextent:      2048s; > CPU:        (none)  Enqlm:      2600  Pgflquo:     100000p > 	 > Thanks,  >  > David Mathog > mathog@seqaxp.bio.caltech.edu @ > Manager, sequence analysis facility, biology division, Caltech   --  D Jilly	- Working from Home in the Chemung River Valley - Lockwood, NY0 	- jilly@clarityconnect.com			- Brett Bodine fan. 	- Mark.Jilson@Compaq.com			- since 1975 or so, 	- http://www.jilly.baka.com               -   ------------------------------  $ Date: Wed, 5 Jul 2000 10:36:34 -0700 From: "Al Gill" <al@spie.org>l* Subject: SCSI hard drive 4 VAX 4700 needed. Message-ID: <sm6sgi3fgas97@corp.supernews.com>  K I'm looking for a 5 to 10GB size SCSI drive to attach to our VAX 4700A.  MynC RF slots are full and I do not want to add another controller, etc.s   ------------------------------   Date: 5 Jul 2000 07:49:34 -0500 9 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)t: Subject: Re: Selling software to corporation AND hobbyists+ Message-ID: <xfrccBtiZUhP@eisner.decus.org>   R In article <Fx75t7.7Ls@spcuna.spc.edu>, Terry Kennedy <terry@gate.tmk.com> writes:    O >   You could also ask if you could hook into the VMS Hobbyist license system -uM > for example, to get a MultiNet PAK you have to provide the checksum of your L > Hobbyist VMS PAK to Process Software (and if the link to the Hobbyist siteO > is down, you don't get a MultiNet PAK [but you may get mail from Hunter apol-:@ > ogizing for the problem 8-], so it's definitely a real check).  F I asked the Hobbyist program folks to include a general "Hobbyist" PAKE that various vendors could honor.  The idea is that there would be noiH way to get that "product" for a commercial site -- one would have to buy4 a vendor-specific license for non-hobbyist machines.  ? The hobbyist organizers said it was a reasonable request, but Ib; gather it has not gotten to the top of their project stack.r   ------------------------------   Date: 5 Jul 2000 07:53:26 -0500w9 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) : Subject: Re: Selling software to corporation AND hobbyists+ Message-ID: <iggZhtITrTfM@eisner.decus.org>   k In article <200007042151_MC2-AB1F-335@compuserve.com>, "Richard B. Gilbert" <DRAGON@compuserve.com> writes:rJ >         Before you spend a lot of time and effort on this, consider thatL > most paging services (around here anyway) can initiate a page via E-mail.=  C But for machines not connected to the Internet, that does not work.i  > There is a lowest-common-denominator wire protocol for dial-up? messages, but it only handles 80 characters when used with Arch  (nee Mobilecomm).U  ? Skytel has a private protocol for which their first approach isT0 to provide object libraries (I bet not for VMS).  C So I think there is a lot of ground that can be covered by software 
 of this sort.    ------------------------------  # Date: Wed, 05 Jul 2000 12:11:31 GMT  From: spolato@my-deja.com : Subject: Re: Selling software to corporation AND hobbyists) Message-ID: <8jv8l9$3au$1@nnrp1.deja.com>   , In article <39626346.E14CD686@videotron.ca>,0   JF Mezei <jfmezei.spamnot@videotron.ca> wrote:G > I intend to eventually productize some software. I would like to makeU it@ > available for purchase/download on the web, with a license key emailed onceA > application is processed (and invoice emailed or paper-mailed).  > G > I think that initially, I will only issue the equivalent of temporary  paks, H > with a permanent one issued when payment is received. This allows easy  > management of free-trials etc.  B I use to codify the name of the system inside the key generated to1 activate the software MenuFinder for OpenVMS/AXP.L@ For example, I use the following Fortran instruction to read the system name:A   STATUS=LIB$GETSYI(SYI$_NODENAME,RIS_NUM,RIS_STRING,LRIS_STRING) C This method avoids to codify in the key the ethernet address of the C network card and the need to regenerating the key when the ethernet  controller is changed.2 The system name is generally more stable, I think.  2 I use to codify an expiration date in the key too.C For temporary use of MenuFinder I generate a key with an expirationdB date of 3 months and when the payment is received I regenerate the0 key with a very long expiration time (40 years).  @ Visit the site :  http://itre.com/mf/download_axp.html to see an example.  
 Sandro Polato  polato@igi.pd.cnr.it    & Sent via Deja.com http://www.deja.com/ Before you buy.c   ------------------------------  # Date: Wed, 05 Jul 2000 15:03:21 GMT*, From: SANFACE Software <sanface@sanface.com> Subject: Shareware: txt2pdf 3.6b) Message-ID: <8jvinc$b17$1@nnrp1.deja.com>   5 txt2pdf is a very flexible and powerful PERL5 programr9 (5 penguins at LinuxBerg, 5 cows at TuCows Mac and BeOS). 5 It's a converter from text files to PDF format files.Q Why do you need txt2pdf?% Most of your documents are text files3F Usually, your reports from legacy applications, DBs, ERP applications, datawarehouse are textual G txt2pdf is a PERL5 script, so you can use it in every operating systemsE$ supported by PERL5 (View the list ofH OS tested) txt2pdf is a native converter, you don't need to pass through PostScript format 2 txt2pdf is specific for text to PDF conversion, soG you can mark yellow, red, green, blue or bold, italic, bolditalic (withi% PERL regular expression) words in theB produced PDF filesH you can produce a 2-columns PDFyou can add page number in every page you can add text at the beginninga and at the end of every file" you can add a border to every page; every word like http://... ftp://... mailto:... https://...:. file:... ldap:... news:...  will become an URL> you can create a link to a specific page within a PDF document http://...pdfdocument#pdfmarkt0 every word like mime:... will become a link that1 launch the correct application and opens the file,9 you can use a few parse commands (e.g.[!blue]...[!/blue])g, you can use background and foreground layers! txt2pdf supports STDIN and STDOUT*  the fee for every licence is $55H SANFACE Software is going to give you a free licence for every good idea or for every good modify   txt2pdf is shareware5 The txt2pdf source code is our company core business.b
 We trust you. $ You can test text2pdf and modify it.E You can't use a modify version of txt2pdf for production purpose. YouX+ can't resell txt2pdf or a modify version ofr* it without SANFACE Software authorization.C You can't copy part of it to include in your source without SANFACEn Software authorization.s   What's new in this version  E withextension feature to don't change the file default extension withr? .pdf (simply txt2pdf will create new PDFs appending .pdf to thed original name)" Lotus Notes note by Charles Daniel) MPEiX note about spoolfiles by Mark Bixbye HP-UX executable version   Test txt2pdf 3.6!i) You can find it at http://www.sanface.com      -- SANFACE Software= Your technology glasses. We help you see your full potential.. http://www.sanface.com mailto:sanface@sanface.com( WAP http://www.sanface.com/wap/index.wml    & Sent via Deja.com http://www.deja.com/ Before you buy.c   ------------------------------  # Date: Wed, 05 Jul 2000 09:44:25 GMTr= From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-)x& Subject: Re: Traceback info from AST ?0 Message-ID: <009EC9CC.8A8E069D@SendSpamHere.ORG>  c In article <110001bfe639$e1e949f0$020a0a0a@xile.realm>, "John E. Malmberg" <wb8tyw@qsl.net> writes: / >JF Mezei <jfmezei.spamnot@videotron.ca> wrote:t >oI >> Is is possible for a routine running in AST mode to know what the main* >program3 >> is up to, or access some traceback information ?e >>M >> (I know how this could manually be acheived, but it would require changingi >allM >> of my code to constantly update some pointer that the AST would use. I was B >> just wondering if this were possible in a more systematic way ? > L >Actually that may be a good idea.  It has lots of possibilities if you want4 >to use a second program to easily spy on the first.  E http://www.tmesis.com/help/@sys$sysdevice:[snapshot]snapshot/snapshot)   ... and it doesn't use an AST!   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM   H        http://www.OpenVMS.digital.com:8000/72final/9996/9996pro_063.htmlH        http://www.OpenVMS.digital.com:8000/72final/9996/9996pro_072.html   ------------------------------  % Date: Wed, 05 Jul 2000 11:04:52 +0100-' From: Adrian Birkett <abirkett@csc.com>4 Subject: Trapping a STOP/IDu' Message-ID: <39630843.A6559095@csc.com>P   All,  C I am currently writing a program that requires an event to be fired8F should the program be terminated for any reason. I can trap a CTRL-Y/CH and a normal process exit using normal system services but all traps are! evaded if someone does a STOP/ID.f  H Is there any way of trapping this or am I wasting my time? Any responses welcome.   Thanks,I   Ade     / ===== My opinion, not that of my employer =====v   ------------------------------  % Date: Wed, 05 Jul 2000 07:14:57 -05002" From: Earl Lakia <lakia@ipact.com> Subject: Re: Trapping a STOP/IDn) Message-ID: <396326C1.119E206D@ipact.com>   E You can write a user written system service for executive mode.  WhenoC the stop occurs, the executive system service can be called as partt; of the image rundown.  That is how the lock manager and RMSWB release locks and close files.  The following shows a snippet from> a system service I wrote.  Of course your image must be linked? to the system service and the system service must be installed.a    extern const PLV user_services = {.#   PLV$C_TYP_CMOD,        /* type */-   0,              /* version */@   {g6     {KERNEL_ROUTINE_COUNT,  /* # of kernel routines */2      EXEC_ROUTINE_COUNT,  /* # of exec routines */1      kernel_table,      /* kernel routine list */ -      exec_table,      /* exec routine list */o3      iqr_rundown,      /* kernel rundown handler */e1      iqr_rundown,      /* exec rundown handler */ *      0,            /* no RMS dispatcher */-      0,            /* kernel routine flags */o+      0}            /* exec routine flags */i   }h }; #ifdef __ALPHA #pragma extern_model restore #endif    /* Rundown handler */   int iqr_rundown()T {O   /* do what ever you need */I   return SS$_NORMAL; }e       -earla --
 Earl D. Lakia 0 Senior Staff Engineer         Web: www.ipact.com4 Snail Mail:                   Email: lakia@ipact.com
 IPACT Inc.1 260 S. Campbell St.           Phone: 219-464-7212  Valparaiso, IN 46383     Adrian Birkett wrote:l   > All, >eE > I am currently writing a program that requires an event to be firedaH > should the program be terminated for any reason. I can trap a CTRL-Y/CJ > and a normal process exit using normal system services but all traps are# > evaded if someone does a STOP/ID.- >-J > Is there any way of trapping this or am I wasting my time? Any responses
 > welcome. >p	 > Thanks,j >: > Ade  > 1 > ===== My opinion, not that of my employer =====0   ------------------------------  % Date: Wed, 05 Jul 2000 07:05:22 -0700 5 From: Richard  <maher_rjNOmaSPAM@hotmail.com.invalid>p Subject: Re: Trapping a STOP/IDh9 Message-ID: <0e24bef6.0e06b187@usw-ex0110-076.remarq.com>    Hi,   ; There is another (better?) example in this news group using:: MACRO. But I don't know how to use deja news (or something9 else?) to search for it. Can someone who knows what to do6) cut and passed it as a reply here please.o  8 I only have a hard copy. I know it's not very big but if; you can save me the typing and assist the author of .0 thens that would be great.   Thanks Richard Maher.r     * Sent from AltaVista http://www.altavista.com Where you can also find related Web Pages, Images, Audios, Videos, News, and Shopping.  Smart is Beautifulr   ------------------------------  % Date: Wed, 05 Jul 2000 08:29:30 -0500u" From: Earl Lakia <lakia@ipact.com> Subject: Re: Trapping a STOP/IDW) Message-ID: <3963383A.662690F8@ipact.com>   > There is a public domain version available from DECUS or me of9 a MACRO written system serivce for OpenVMS/VAX.  However,:6 I don't  use MACRO much on the Alpha platform when one4 can use C.   There is a good document in the OpenVMSA documentation that deos a good job exlpaining how to write a user : written system service.  It describes the rundown methods.   -earlv lakia@ipact.comi   Richard wrote:   > Hi,i > = > There is another (better?) example in this news group using.< > MACRO. But I don't know how to use deja news (or something; > else?) to search for it. Can someone who knows what to dox+ > cut and passed it as a reply here please.l >c: > I only have a hard copy. I know it's not very big but if= > you can save me the typing and assist the author of .0 theno > that would be great. >b > Thanks Richard Maher.d >o > * Sent from AltaVista http://www.altavista.com Where you can also find related Web Pages, Images, Audios, Videos, News, and Shopping.  Smart is Beautiful.   ------------------------------  % Date: Wed, 05 Jul 2000 15:54:46 +0100 - From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>n Subject: Re: Trapping a STOP/ID() Message-ID: <39634C36.8B797C74@bbc.co.uk>b   Adrian Birkett wrote:;   > All, >rE > I am currently writing a program that requires an event to be firedcH > should the program be terminated for any reason. I can trap a CTRL-Y/CJ > and a normal process exit using normal system services but all traps are# > evaded if someone does a STOP/ID.0 >uJ > Is there any way of trapping this or am I wasting my time? Any responses
 > welcome. >t   Adrian  A If you are creating a separate detached process you can specify ac termination mailboxe3 that is sent a message when the process terminates.S     ie   RUNg  	   Processe       /MAILBOX             /MAILBOX=unit   F        Specifies the unit number of a mailbox to receive a terminationD        message when the created process is deleted. If no mailbox isH        specified, the creating process receives no notification when the7        subprocess or detached process has been deleted.        RUN Process Subtopic?T
 RUN Subtopic?   > or the docs for the $CREPRE system service (mbxunt arguement).  G It's been a while since I used it but I'm sure you get a signal even on  stop/id.  L The other thing to do is use the lock manager and in the program exclusively  L take out a user mode lock. Then, in the monitoring program, request the lock and L specify an AST or event flag to be set when it is available. Again, its been  I a while, the exact terminology has escaped me, but either approach should  work, G the lock manager one is more flexible in that a detached process is notn	 required.p   HTH.     --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: Wed, 05 Jul 2000 11:33:26 -0500i" From: Earl Lakia <lakia@ipact.com> Subject: Re: Trapping a STOP/IDt) Message-ID: <39636356.92C5BDA1@ipact.com>e  = Tim's Solution are good.  I used  a simple foriegn command toi; to this with the lock manager and a DCL script (used it for ? automatically switching applications on a cluster when a system C faulted).  I've also used the mailbox method that Tim stated, againT? a good solution.  However, with both, you need a second process 8 to do the cleanup.  This can sometimes can be a problem.   -earl    Tim Llewellyn wrote:   > Adrian Birkett wrote:  >  > > All, > >UG > > I am currently writing a program that requires an event to be fireddJ > > should the program be terminated for any reason. I can trap a CTRL-Y/CL > > and a normal process exit using normal system services but all traps are% > > evaded if someone does a STOP/ID.  > >1L > > Is there any way of trapping this or am I wasting my time? Any responses > > welcome. > >  >  > Adrian > C > If you are creating a separate detached process you can specify a  > termination mailbox 5 > that is sent a message when the process terminates.  >  > ie >l > RUN  >0 >   Process  >  >     /MAILBOX >  >           /MAILBOX=unitS >:H >        Specifies the unit number of a mailbox to receive a terminationF >        message when the created process is deleted. If no mailbox isJ >        specified, the creating process receives no notification when the9 >        subprocess or detached process has been deleted.  >i > RUN Process Subtopic?n > RUN Subtopic?r >l@ > or the docs for the $CREPRE system service (mbxunt arguement). > I > It's been a while since I used it but I'm sure you get a signal even onm
 > stop/id. >-N > The other thing to do is use the lock manager and in the program exclusively >AN > take out a user mode lock. Then, in the monitoring program, request the lock > and3N > specify an AST or event flag to be set when it is available. Again, its been > K > a while, the exact terminology has escaped me, but either approach shouldn > work,eI > the lock manager one is more flexible in that a detached process is not  > required.e >i > HTH. >o > --8 > Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project2 > MedAS at the BBC, Whiteladies Road, Bristol, UK.C > Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uke >dC > I speak for myself only and my views in no way represent those ofl > MedAS or the BBC.o   ------------------------------  $ Date: Wed, 5 Jul 2000 11:40:29 -0400- From: "Peter Weaver" <peter.weaver@stelco.ca>P6 Subject: Upgrade to VMS 7.2 fails on Help Message File. Message-ID: <sm6m4j8egas65@corp.supernews.com>  @ I am upgrading my Hobbyist Charon-VAX from V6.2 to V7.2 and haveC placed the Help message on a different disk. After answering all offD the questions and rebooting the node I see the following (the system8 drive is DKA0, the drive for the help messages is DKA1);  *     Restoring OpenVMS library save set ...1 %BACKUP-I-STARTVERIFY, starting verification passb  +     Restoring OpenVMS optional save set ...i1 %BACKUP-I-STARTVERIFY, starting verification passo0 %MOUNT-I-MOUNTED, DISK1 mounted on _WEAVER$DKA1:  /     Restoring OpenVMS Help Message database ...:1 %BACKUP-I-STARTVERIFY, starting verification passa< %SET-F-SEARCHFAIL, error searching for DKA1:[HELP_MESSAGE].; -RMS-E-FNF, file not found, %MOUNT-F-DEVMOUNT, device is already mounted8   SYSTEM       job terminated at  5-JUL-2000 07:16:58.31     Accounting information:eD   Buffered I/O count:            7397         Peak working set size: 2021B   Direct I/O count:             11405         Peak page file size: 6021>   Page faults:                   4980         Mounted volumes: 2tA   Charged CPU time:           0 00:20:25.05   Elapsed time:     0  00:27:12.16d  # Any idea on what is happening here?0   -- Peter Weaver   ------------------------------    Date: 05 Jul 2000 10:18:32 +0200G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>r Subject: Re: VAX on Intel?H Message-ID: <y43dlpxa5j.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  3 Glen Martin <GLENMARK@utxvms.cc.utexas.edu> writes::  & > > So why didn't you go for OpenMail?O > Because we're not an HPUX shop? (I work in our University's OpenVMS/NT Group)$  N OpenMail has got nothing to do with HPUX. It is also available for WNT and for
 Linux/Intel. i  N > Just as well. I've heard rumors that OpenMail is even flakier than Exchange.  . I've heard quite the opposite, including here.   	Jan   ------------------------------  % Date: Wed, 05 Jul 2000 12:27:17 -0400c# From: John Vottero <John@MVPSI.com> ) Subject: RE: VMS can tail -f ? if so how? D Message-ID: <C15945A9D9EFCF11BA8B08002BBF1CCC0CD74F@berry.mvpsi.com>  K It depends upon how you establish the connection.  It could be as simple as I specifying keepalives in your service definition.  For Compaq TCP/IP (akas  UCX) it would be something like:  8 $ TCPIP SET SERVICE yourservice/SOCKET_OPTIONS=KEEPALIVE  G The client could also request keepalives with the setsockopt() routine.G   > -----Original Message-----B > From: tony_barratt@my-deja.com [mailto:tony_barratt@my-deja.com]& > Sent: Tuesday, July 04, 2000 5:24 AM > To: Info-VAX@Mvb.Saic.Com + > Subject: RE: VMS can tail -f ? if so how?t >  > - > Well sounds good. PLease, how do I do that? F > In article <C15945A9D9EFCF11BA8B08002BBF1CCC0CD74A@berry.mvpsi.com>,( >   John Vottero <John@MVPSI.com> wrote:H > > Why don't you just use keepalives when you establish the connection? > >y  > > > -----Original Message-----F > > > From: tony_barratt@my-deja.com [mailto:tony_barratt@my-deja.com]) > > > Sent: Monday, July 03, 2000 7:07 PMt > > > To: Info-VAX@Mvb.Saic.Comy/ > > > Subject: Re: VMS can tail -f ? if so how?, > > >  > > >:B > > > I was wondering is there a way to create an repetative event > > > which will > > > appear in the alert log?# > > > I'd like to have a line like:  > > > hostname count timestamp > > > every 3 secs, say.H > > > So I'll know if the network connectivity goes away, and silence is > noB > > > longer a guarantee of a healthy VMS system (because I have a
 > watchdogG > > > on the remote host that notices when the repetative events stop). F > > > Is there a fine manual I could read that might explain how to do > this > > > in DCL(?).A > > > Or could some kind VMS person give a visitor from unix-landa > > > some sample  > > > code :-) > > >u	 > > > TIAi > > >a
 > > > Tony > > >i/ > > > In article <8jqg2k$pas$1@nnrp1.deja.com>,y' > > >   tony_barratt@my-deja.com wrote:b$ > > > > Ah yes well take your point.< > > > > Basically I want to monitor a couple of pairs of VMS > > > machines plus the # > > > > application sitting on top.tC > > > > We monitor quite a few different platforms already, but ther
 > approach > > > isF > > > > the same. Persuade system to append alerts/error msgs to a log > file,.D > > > > and squirt added lines down a socket to the fault management > system	 > > > for/ > > > > parsing.F > > > > I have in mind, after interogating some VMS savvy peeps, using > > > DECevent: > > > > which is, so I believe, the interface between the  > system user andg > > > VMS's  > > > > event logging system.s7 > > > > It is invoked with the DIAGNOSE command and if i > specified with the@ > > > > /CONTINUOUS  parameter , events are reported as they are > > > logged by theiC > > > > system so in other words, events are viewed in real time. Ar	 > furtherbG > > > > parameter is /OUTPUT  in which I can specify an output file foro > ther  > > > > events to be written to. > > > >X > > > > OK so far?B > > > > So I reckon a type/tail/continuous will push the data over > > > ethernet top$ > > > > my waiting info hungry host? > > > >F8 > > > > This does seem a bit unix-like. Is there a more  > VMS-like way to do > > > this?t > > > >h+ > > > > BTW Thanks for the rapido response.t > > > >l > > > > Tony > > > >d5 > > > > In article <3960A1E1.228D73AF@earthlink.net>,hB > > > >   "David J. Dachtera" <djesys.nospam@earthlink.net> wrote:) > > > > > tony_barratt@my-deja.com wrote:a > > > > > >h > > > > > > Hi Guys,? > > > > > > I have a task set me, to push the text appended to s > a log file > > > down aA > > > > > > socket for processing (by a fault management system).nB > > > > > > To explain further, in unix, with which I am familiar, > > > we set up a ; > > > > > > service, associated with a udp port number, in 4 > /etc/services, > > > which,D > > > > > > when we do a 'telnet machine portnumber' invokes a shell > script
 > > > that? > > > > > > invokes a 'tail -f /path/logfilename' which allows 9
 > the logfilea > > > > apendsF > > > > > > to appear on the remote machine (that invoked the telnet).? > > > > > > I gather that VMS's Digital TCP/IP has +/- the samen > > > functionality  > > > as) > > > > > > for example Solaris's TCP/IP. D > > > > > > Can some kind person therefore advise on a way to <unix- > > > speak>tail a( > > > > > > file </unix-speak> with VMS?	 > > > > > G > > > > > I find myself wondering if there isn't a better way to do thet > same? > > > > > thing. What is that you're trying to monitor? ...and   > what needs > to' > > > > > receive the monitored output? 	 > > > > >oH > > > > > Doing UN*X-ly things on VMS is the "politically correct" thing > to	 > > > do,pF > > > > > but tends to be costly to develop and maintain, and tends to > lack > > > > > reliability.	 > > > > >a@ > > > > > Doing VMS things the VMS way, by contrast, tends to be > > > more stable, > > > > moreE > > > > > reliable, and less costly in the long run (less development  > time,-
 > > > less > > > > > maintenance).0	 > > > > >1 > > > > > -- > > > > > David J. Dachterae > > > > > dba DJE Systemsp  > > > > > http://www.djesys.com/	 > > > > >5D > > > > > Unofficial Affordable OpenVMS Home Page and Message Board:) > > > > > http://www.djesys.com/vms/soho/ 	 > > > > >f > > > >a. > > > > Sent via Deja.com http://www.deja.com/ > > > > Before you buy.  > > > >b > > >t > > >e, > > > Sent via Deja.com http://www.deja.com/ > > > Before you buy.t > > >a > >a > >  >  > ( > Sent via Deja.com http://www.deja.com/ > Before you buy.i >    ------------------------------  # Date: Wed, 05 Jul 2000 16:19:17 GMTe From: bawilhelm@my-deja.com0$ Subject: Re: VMS Printing using PCL6) Message-ID: <8jvn5g$ej6$1@nnrp1.deja.com>g   Barry,  E We do a lot of embedding of PCL5 in print jobs from VMS.  We just got A some new printers that support PCL6.  I was curious if you have aeC complete command reference for PCL6 and if so where did you get it.l   Thanks  
 Brent Wilhelms    & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  # Date: Wed, 05 Jul 2000 14:16:40 GMTs From: hanghien@my-deja.com% Subject: What's wrong with F$LOCATE ?n) Message-ID: <8jvfvm$8ph$1@nnrp1.deja.com>D   $  sh   log  hxn$tools  ?    "HXN$TOOLS" = "HXN$DISK:[HA$NG01.TOOLS]" (LNM$PROCESS_TABLE)#   $  write sys$output    f$locateO ("[","hxn$tools")o 9x   $  write sys$output    f$locate  ("]","hxn$tools")n 9i    & Sent via Deja.com http://www.deja.com/ Before you buy.i   ------------------------------  % Date: Wed, 05 Jul 2000 15:54:35 +0100x- From: "POWERS, John" <John.POWERS@sema.co.uk>o) Subject: RE: What's wrong with F$LOCATE ?e; Message-ID: <D30A62ABC710D211AEE100A0C9D615EE3163FB@REAES2>a  ! Well 2 points to make to this onea  : > From: hanghien@my-deja.com [mailto:hanghien@my-deja.com] > Sent: 05 July 2000 15:17 > To: Info-VAX@Mvb.Saic.Com ' > Subject: What's wrong with F$LOCATE ?  - Nothing, it's great!   >  >  > $  sh   log  hxn$tools > A >    "HXN$TOOLS" = "HXN$DISK:[HA$NG01.TOOLS]" (LNM$PROCESS_TABLE)  > ! > $  write sys$output    f$locate9 > ("[",) > 9p > ! > $  write sys$output    f$locateo > ("]","hxn$tools")  > 9c > C Firstly, check the documentation. f$locate requires a string, not atH logical name. You can use a quoted string, a symbol, or the return valueE from another lexical. The way you have done it, it will look for the d@ square brackets in the string "hxn$tools", and there ain't none.  C If you really want this information, you will need to translate thejD logical name into a string. Look up the lexical f$trnlnm, which will do this.  e.g.7 $ write sys$output f$locate ("[",F$TRNLNM("hxn$tools"))   H Secondly, do you really need this information? If you are trying to findG the wherabouts of the square brackets in order to extract the directory G information (the only reason I could think of to get this information), C then there is a much easier way. Look up the lexical F$PARSE - e.g.   5 $ WRITE SYS$OUTPUT F$PARSE("HXN$TOOLS",,,"directory")  [HA$NG01.TOOLS]p   HTH     -- Cheers, John  F  - Note  This message represents my opinions and nothing else, not theI   opinion of SEMA, my family, or the cricket club - though my dog Meg diddE   nod in agreement whilst I was typing. If you have any problems thencD   please complain to her (or me, but not SEMA, my family or the CC).  g    K ___________________________________________________________________________0B This email is confidential and intended solely for the use of the H individual to whom it is addressed. Any views or opinions presented are E solely those of the author and do not necessarily represent those of   Sema Group. M If you are not the intended recipient, be advised that you have received thisnI email in error and that any use, dissemination, forwarding, printing, or h- copying of this email is strictly prohibited.e  E If you have received this email in error please notify the Sema Group . Helpdesk by telephone on +44 (0) 121 627 5600.K ___________________________________________________________________________t   ------------------------------  # Date: Wed, 05 Jul 2000 13:54:27 GMTi. From: Michael Austin <maustin@nc.prestige.net>) Subject: Re: What's wrong with F$LOCATE ? / Message-ID: <39634BB4.75EA891C@nc.prestige.net>u  , This is a multi-part message in MIME format.& --------------5FB469130AFAF58CAFDFD11E* Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bitt  G your problem is that you cannot use a logical in this context, f$locates& only works on symbols or text strings.   try:9 $ write sys$output    f$locate("[",f$trnlnm("hxn$tools"))4 9 9 $ write sys$output    f$locate("]",f$trnlnm("hxn$tools"))j 23   Michael Austin DBA Consultant   hanghien@my-deja.com wrote:k   > $  sh   log  hxn$tools > A >    "HXN$TOOLS" = "HXN$DISK:[HA$NG01.TOOLS]" (LNM$PROCESS_TABLE)n >o! > $  write sys$output    f$locate  > ("[","hxn$tools")a > 9I >n! > $  write sys$output    f$locateo > ("]","hxn$tools")l > 9e >e( > Sent via Deja.com http://www.deja.com/ > Before you buy.0  & --------------5FB469130AFAF58CAFDFD11E- Content-Type: text/x-vcard; charset=us-ascii;m  name="maustin.vcf"f Content-Transfer-Encoding: 7bitc, Content-Description: Card for Michael Austin  Content-Disposition: attachment;  filename="maustin.vcf"S   begin:vcard  n:Austin;Michael   tel;work:704-947-1089  x-mozilla-html:FALSE org:Michael Austin, Incb
 adr:;;;;;; version:2.1e+ email;internet:michaelaustininc@hotmail.come title:Presidentt x-mozilla-cpt:;0 fn:Michael E. Austin	 end:vcardg  ( --------------5FB469130AFAF58CAFDFD11E--   ------------------------------  % Date: Wed, 05 Jul 2000 16:03:22 +0100 # From: Kevin B Black <kbb@ceh.ac.uk>d) Subject: Re: What's wrong with F$LOCATE ?a) Message-ID: <s9635c6f.001@wpo.nerc.ac.uk>d  A Nothing is wrong with it. F$LOCATE returns the offset (zero basedn@ index) of the substring being searched for, or the length of theE string being searched if the search substring was not found. F$LOCATE + does not translate logical names. Thereforee  # $ write sys$output("Z","ABCDEFGHI")e 9 # $ write sys$output("A","ABCDEFGHI")e 0p $ show log XDPS$EXAMPLES8    "XDPS$EXAMPLES" = "SYS$COMMON:[SYSHLP.EXAMPLES.XDPS]" (DECW$LOGICAL_NAMES)0 $ write sys$output f$locate("[","XDPS$EXAMPLES") 13: $ write sys$output f$locate("[",f$trnlnm("XDPS$EXAMPLES")) 11  @ Therefore I would suggest, to get the results I guess you may be expecting, you wanto  5 $ write sysoutput f$locate("[",f$trnlnm("hxn$tools"))p   regards KevinBB.  3 >>> <hanghien@my-deja.com> 05-Jul-2000 15:16:40 >>>a $  sh   log  hxn$tools  ?    "HXN$TOOLS" = "HXN$DISK:[HA$NG01.TOOLS]" (LNM$PROCESS_TABLE)p   $  write sys$output    f$locate5 ("[","hxn$tools")m 9    $  write sys$output    f$locate  ("]","hxn$tools")m 9d    ' Sent via Deja.com http://www.deja.com/   Before you buy.o   ------------------------------   End of INFO-VAX 2000.373 ************************