1 INFO-VAX	Fri, 28 Apr 2000	Volume 2000 : Issue 237       Contents:, Re: Any ALPHAbook (Tadpole) users out there?, Re: Any ALPHAbook (Tadpole) users out there?# Backup /verify causes parity errors ' Re: Backup /verify causes parity errors $ Re: Backup Strategies for Enterprise) Re: Calculating Discordian dates from DCL % Re: Can Windows NT connect to HSZ80s? % Re: Compaq still seen as a PC company  Re: Console Firmware From VMS  Current VMS Perl availability  DCPS and TI Microlaser Pro E  Disk I/O Statistics Don't Add Up% Re: Disk subsystems for OpenVMS Alpha % Re: Disk subsystems for OpenVMS Alpha % Re: Disk subsystems for OpenVMS Alpha   DLT Drive Software Write Locking7 Re: Does backup command backup used or allocated space? ! Re: Dropping DECnet..don't do it! $ Re: F$File_Attribute from a program?$ Re: F$File_Attribute from a program?$ Re: F$File_Attribute from a program?$ Re: F$File_Attribute from a program?$ Re: F$File_Attribute from a program?! File Transfer Management software @ Re: Including CD media with systems? (was Re: OpenVMS marketing)@ Re: Including CD media with systems? (was Re: OpenVMS marketing) Re: Ingres and DefragmentationB Re: Is it possible to get rid of this annoying boot-up message ???? Re: Multiport serial card on PC as AlphaServer serial consoles?  Re: OpenML for OpenVMS? + Re: OpenVMS and WNT integration for Dummies + Re: OpenVMS and WNT integration for Dummies  Re: OpenVMS marketing + Re: Porting To VMS: select function problem + Re: Porting To VMS: select function problem  Print queue -> file  Re: Print queue -> file  Re: Print queue -> file  Re: queue status values  Re: So who will buy VMS ?  Re: SRM documentation $ Re: Sun is no longer the dot in .com& Re: TCPIP V5.0A configuration problems& Re: TCPIP V5.0A configuration problems& RE: TCPIP V5.0A configuration problems& Re: TCPIP V5.0A configuration problems Time Sync'ing a VMS server Re: Time Sync'ing a VMS server Re: Time Sync'ing a VMS server* Re: TPU/EVE : decwindows vs character cell Re: USERNAME Re: Verify of Backups 3 VMS/UCX responds to UDP broadcast with ICMP message 7 Re: VMS/UCX responds to UDP broadcast with ICMP message 7 Re: VMS/UCX responds to UDP broadcast with ICMP message 7 Re: VMS/UCX responds to UDP broadcast with ICMP message 7 RE: VMS/UCX responds to UDP broadcast with ICMP message 7 RE: VMS/UCX responds to UDP broadcast with ICMP message 1 Re: Where is the lexical f$getjpi("","Connectim") 1 RE: Where is the lexical f$getjpi("","Connectim") 9 Re: Why does Pathworks(Mac) change mod date of .dir file? < [upd]Re: Does backup command backup used or allocated space?  F ----------------------------------------------------------------------  % Date: Fri, 28 Apr 2000 09:14:58 +0100 9 From: "Colin Butcher" <colin.butcher@xdelta.co.uk.nospam> 5 Subject: Re: Any ALPHAbook (Tadpole) users out there? A Message-ID: <956909674.12545.0.nnrp-04.9e98f8e9@news.demon.co.uk>   < "Skipper W. Morris" <morris@iridium.mv.net> wrote in message" news:8ea92c$j54$1@pyrite.mv.net...G > I've just started playing with a Tadpole (Alpha notebook) that's been E > sitting around work here unused.  I've run into a few problems that ? > hopefully someone already has seen and knows then answer for.  >   K Great machines for training / demo use. I had two for a while (on loan from 4 Tadpole) and would really like to get one of my own.  J > 1. UCX won't configure.  The ethernet device "EO:" is not in the list ofH > network devices in UCX$CONFIG.COM.  Is it safe to hack the config 'com1 > file?  Or do I need a different version of UCX?  >   , Try UCX DEFINE COMMUNICATION_CONTROLLER EO -<                 /INTERFACE_TYPE=O /TYPE=(ETHERNET,CLUSTER) -(                 /DESCRIPTION=PCMCIA_3COM  L That's from the "Alphabook1 startup guide". Sounds like you're on VMS 6.2 orG 7.1 with UCX 4.2 - try upgrading to VMS 7.2-1 and TCP/IP services V5.0A   C > 2. I can't do a windows login because of a missing license.  I've D > installed all the normal ones: OPENVMS-ALPHA, VAX-VMS, VAXCLUSTER,D > VMSCLUSTER, DVNETRTG, DVNETEXT, DW-MOTIF, UCX.  What am I missing? >   C I don't see OPENVMS-ALPHA-USER in there. All I think you need for a L stand-alone machine is OPENVMS-ALPHA, OPENVMS-ALPHA-USER, DVNETEND, DW-MOTIF and UCX.  G > 3. Anyone know where documentation, etc, might be located on the Web?  > I haven't found any so far.  >  > thanks > /Skip   I I have a paper doc set for the AlphaBook1 (Startup guide and User guide), F but no machine to go with it. Contact me directly if you would like toK borrow it. Maybe you could convert it to PDF and put it up on the web (with K Tadpole's permission of course). Meantime I'll ask some people I know there : in case it's still available in a machine readable format.  3 Be aware that you will also need to make changes to G DECW$PRIVATE_SERVER_SETUP.COM to set display resolution and to enable / B disable the external video out. Again it's in the "startup guide".  : $ if (f$locate("ALPHABOOK1",hwname) .ne. f$length(hwname)) $ then! $     decw$xsize_in_pixels == 800 ! $     decw$ysize_in_pixels == 600 % $     decw$server_display_select == 3 L $     say "... Screen Size forced to 800x600 with both LCD and CRT enabled." $ endif   H I remember being badly bitten by the larger (2Gb?) SCSI discs and eitherI OpenVMS V6.2 or V7.1 - there is a SCSI devices patch you have to install, J otherwise the discs gradually become corrupted with continued use. I can'tI access the details right now as my VMS machine is temporarily dismantled.   
 Cheers, Colin   1 Email (munged): colin.butcher@xdelta.co.uk.nospam    ------------------------------   Date: 28 Apr 2000 08:44:05 GMT3 From: gartmann@immunbio.mpg.de (Christoph Gartmann) 5 Subject: Re: Any ALPHAbook (Tadpole) users out there? 0 Message-ID: <8ebj0l$ldv$1@n.ruf.uni-freiburg.de>  Z In article <8ea92c$j54$1@pyrite.mv.net>, morris@iridium.mv.net (Skipper W. Morris) writes:F >I've just started playing with a Tadpole (Alpha notebook) that's beenD >sitting around work here unused.  I've run into a few problems that> >hopefully someone already has seen and knows then answer for. > I >1. UCX won't configure.  The ethernet device "EO:" is not in the list of G >network devices in UCX$CONFIG.COM.  Is it safe to hack the config 'com 0 >file?  Or do I need a different version of UCX?  # Here the ethernet device is EOA0: .   B >2. I can't do a windows login because of a missing license.  I'veC >installed all the normal ones: OPENVMS-ALPHA, VAX-VMS, VAXCLUSTER, C >VMSCLUSTER, DVNETRTG, DVNETEXT, DW-MOTIF, UCX.  What am I missing?   4 What about OPENVMS-ALPHA-USER and eventually OPEN3D?  F >3. Anyone know where documentation, etc, might be located on the Web? >I haven't found any so far.  L I don't think there is any information on the web. In fact, there isn't muchK information in the docs either. On the other hand, it is a real VMS system, % so the usual VMS information applies.    Regards,    Christoph Gartmann   H -----------------------------------------------------------------------+H | Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452 |H | Immunbiologie                                                        |H | Postfach 1169                 Internet: gartmann@immunbio.mpg.de     |H | D-79011  Freiburg, FRG                                               |H +------------ http://www.immunbio.mpg.de/english/menue.html -----------+   ------------------------------   Date: 26 Apr 2000 10:37:47 GMT3 From: "Tomekd" <Tomasz.Dryjanski@intl.fritolay.com> , Subject: Backup /verify causes parity errors5 Message-ID: <01bfaf6b$52901900$25e1c6a5@8637bbd40750>    Hi all!   J Do you have any idea why backup causes parity errors when run with /verifyJ qualifier? The problem occurs two - three days after booting, during night backup procedure. I We have tried using cleaning tapes, using new tapes, replacing the drive, E unplugging and plugging again all SCSI connectors, replacing the SCSI $ terminator - and there is no result.   Here are the details:  AlphaServer 1000 4/2665 OpenVMS V7.1-1H1 with all mandatory pathes up-to-date H TLZ07-BA - internal, TLZ07-DA - external (we are sure the drive is good)# served by on-board SCSI controller.   , And there is a typical errorlog entry below:  *  HW_MODEL: 000004F0 Hardware Model = 1264.  $  DEVICE ERROR AlphaServer 1000 4/266    TLZ07 SUB-SYSTEM           HW REVISION     30454234 9                                        HW REVISION = 4BE0         ERROR TYPE            05 C                                        EXTENDED SENSE DATA RECEIVED         PORT STATUS     00000001 J                                        %SYSTEM-S-NORMAL, NORMAL SUCCESSFUL2                                         COMPLETION        SCSI CMD        FE000008                             0000 +                                        READ         SCSI STATUS           02 6                                        CHECK CONDITION    EXTENDED SENSE DATA          EXTENDED SENSE  00030070                         0A000000                         00000000                         46000031                             0000 3                                        MEDIUM ERROR >                                        MEDIUM FORMAT CORRUPTED        UCB$L_ERTCNT    00000010 <                                        16. RETRIES REMAINING        UCB$L_ERTMAX    00000010 <                                        16. RETRIES ALLOWABLE        ORB$L_OWNER     00400086 :                                        OWNER UIC [100,206]        UCB$L_CHAR      0DCC4021 6                                        RECORD ORIENTEDB                                        "SEQUENTIAL BLOCK" ORIENTED4                                        FILE ORIENTED0                                        AVAILABLE.                                        MOUNTED4                                        ERROR LOGGING0                                        ALLOCATED.                                        FOREIGN7                                        CAPABLE OF INPUT 8                                        CAPABLE OF OUTPUT        UCB$L_STS       00001910 -                                        ONLINE +                                        BUSY 5                                        SOFTWARE VALID 9                                        UNLOAD AT DISMOUNT         UCB$L_OPCNT     000A03A1 >                                        656289. QIO'S THIS UNIT        UCB$L_ERRCNT    00000001 :                                        1. ERRORS THIS UNIT        IRP$L_BCNT      0000FE00 C                                        TRANSFER SIZE 65024. BYTE(S)         IRP$L_BOFF      00000000 <                                        TRANSFER PAGE ALIGNED        IRP$L_PID       00160080 6                                        REQUESTOR "PID"        IRP$Q_IOSB      00000000 C                        00000000        IOSB, 0. BYTE(S) TRANSFERRED    Thanks in advance for any help!    Tomek D.   ------------------------------  % Date: Fri, 28 Apr 2000 07:52:47 -0700 0 From: "William S. LaCounte" <vmsmanager@ups.edu>0 Subject: Re: Backup /verify causes parity errors# Message-ID: <3909A5BF.2B94@ups.edu>   H I have received the same "extended sense error" on my DPW 433au at home,, except I get the error when mounting a disk.; My system disk is a RZ1CB which mounts fine. My second disk . is a RZ1CC which gives the error during mount.  F I suspect the SCSI protocol in the RZ1CC is at a higher level then the# SCSI driver on my system (VMS 7.2).   2 Perhaps you have the same problem with your TLZ07.   Bill   ------------------------------  % Date: Fri, 28 Apr 2000 16:06:53 +0200 $ From: Paul Sture <sture.ch@sture.ch>- Subject: Re: Backup Strategies for Enterprise + Message-ID: <VA.0000001c.16b2796b@sture.ch>   > In article <8e4qe6$a6j$1@cyw.uklinux.net>, Dave Pickles wrote:/ > From: Dave Pickles <davep@nugate.demon.co.uk>  > Newsgroups: comp.os.vms / > Subject: Re: Backup Strategies for Enterprise % > Date: Tue, 25 Apr 2000 20:13:24 GMT  >  > Paul Sture wrote:  > {stuff about NT backups}L > > There is another solution, to be found at www.ultrabac.com. Not only is M > > this the fastest NT backup I tried, but reliable, fast, and would backup  M > > to removable disk too. And it can boot some form of standalone system to  K > > do a restore. Not cheap, but I got good recommendations from those who  A > > bought it (and they actually did a native Alpha version too).  > J > The "Image" backup tool (actually an analogue of VMS's Standalone BackupI > /PHYSICAL) is zero-cost. And the paid-for product is the best NT backup  > solution I've found. > K Thanks for that info. I see it also works for W98 nowadays, so I'll try it   out.  N On related note, their website also describes a new feature which can be used M to get to savesets on DLTs quickly to ensure quick restores (via an index on  N disk). Hmm - wonder if that or a similar concept could be used to enhance VMS  restores...  ___ 
 Paul Sture Switzerland    ------------------------------  # Date: Fri, 28 Apr 2000 08:13:39 GMT  From: graculuss@my-deja.com 2 Subject: Re: Calculating Discordian dates from DCL) Message-ID: <8ebh7a$64d$1@nnrp1.deja.com>    Emergency bug fix:  	 Details:- F        The season was being incorrectly changed one day early, leading> to anomalies such as the day after 72nd of Discord was 73rd of# Confusion (and it _WAS_ confusing).    Fix:- G        Use any suitable editor (the more adventurous will use Teco, and E the really weird will copy it to a Micro$oft box and use WORD - yes I 9 do know someone who does this !!!!) and replace the line:    $    DMonthnum = (Jdate/73)    with   $    DMonthnum = ((Jdate-1)/73)   C Luckily this bug was reported early by one of my users otherwise we E could have seen a meltdown in the financial markets the like of which E hasn't been seen since Gradnock Gruntfuttock's calculator display was 	 inverted.   @ Rumours that this bug is responsible for the recent falls in the? Technology market sectors are vigoruously denied by the author.     4 In article <38FE1596.FB176413@netscapeonline.co.uk>,   mpatt@bigfoot.com wrote: > Hi, < > 	Working in the finance sector I keep getting asked for th    & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  # Date: Fri, 28 Apr 2000 15:03:19 GMT % From: Alan Greig <agreig@my-deja.com> . Subject: Re: Can Windows NT connect to HSZ80s?) Message-ID: <8ec97a$vss$1@nnrp1.deja.com>   4 In article <3908E287.B13A4998@ameritech.net.nospam>,2   Scott Vieth <svieth@ameritech.net.nospam> wrote: > Hi:  > D > Do the HSZ80-based storage arrays support Windows NT servers or do> > I need to stick with the HSZ70-based line for my NT project?    @ The HSZ80 most definitely supports NT. Even allows NT to coexist@ on the same SCSI bus as VMS systems if you are very, very brave. --
 Alan Greig    & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  % Date: Fri, 28 Apr 2000 08:19:00 -0500 1 From: "Dave Gudewicz" <david.gudewicz@abbott.com> . Subject: Re: Compaq still seen as a PC company8 Message-ID: <8ec2ts$gdk$1@fizban.fizban.pprd.abbott.com>  L OTOH.  Here's a link that does "suggest" that Compaq may be shortly changing its tune.  We'll see.     L http://www.cnetinvestor.com/newsitem-bloomberg.asp?symbol=85547212&Ticker=CP Q    Dave...   F Wayne Sewell <wayne@tachyon.xxx.070701.killspam.0199> wrote in message( news:WzsxKq1pgO3O@tachxxsoftxxconsult...: > In article <3907B82F.8654D7B7@vl.videotron.ca>, JF Mezei) <jfmezei.spamnot@vl.videotron.ca> writes: K > > Read a report on the "turning around" at Compaq following its financial  > > results release. > >  > > Article at: B http://dailynews.yahoo.com/h/nm/20000426/tc/compaq_analysts_1.html > > J > > Article mentions only Wintel operations and how Compaq's fortunes will turnI > > around with Windows 2000 sales as businesses upgrade their servers to  run H > > Win2k. No mention of Alpha, Unix or VMS at all in the article. (As a	 matter of ' > > fact, no mention of Tandem either).  >  > K > Given the lack of priority-for/promotion-of vms and alpha, it seems as if , > compaq sees *itself* as a pc company only. >  > -- > L ============================================================================ ===o: > Wayne Sewell, Tachyon Software Consulting  (281)812-0738 wayne@tachyon.xxxh8 > http://www.tachyon.xxx/www/tachyon.html and wayne.htmlH > change .xxx to .com in addresses above, assuming you are not a spambot :-)a >pL ============================================================================ ===dK > Butler:"Gentlemen!"  Curly(as he and other Stooges look around):"Who camel in?"   ------------------------------  % Date: Fri, 28 Apr 2000 08:04:13 -0500E1 From: "Dave Gudewicz" <david.gudewicz@abbott.com>t& Subject: Re: Console Firmware From VMS8 Message-ID: <8ec225$g7p$1@fizban.fizban.pprd.abbott.com>  K It depends.  I know I didn't enter the serial number in SRM for a couple ofaI new DS20Es we have here and show cpu/full shows the serial number.  OtherfG systems do not.  Maybe newer systems are having this done when they aree built.  Dunno for sure.i   Dave...e  J > Did you enter the serial number in the SRM console ? I did, and show cpuE > /full does show me the serial number now, but you have to enter thek, > serial number manually in the SRM console. >  > >p > >  --k: > > Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project4 > > MedAS at the BBC, Whiteladies Road, Bristol, UK.E > > Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.ukf > >eE > > I speak for myself only and my views in no way represent those ofN > > MedAS or the BBC.e   ------------------------------  % Date: Fri, 28 Apr 2000 09:49:39 -0700e5 From: coyote1 <coyote1NOcoSPAM@flashmail.com.invalid> & Subject: Current VMS Perl availability9 Message-ID: <0ecf8657.d4a32b31@usw-ex0101-005.remarq.com>w  ; What is the current version of Perl that runs on VMS (7.1)?e  = I took a look thru the list archives, but didn't see anything5: about Perl 5.6.  I downloaded the source kit, but it bombs> miserably once it starts to compile (DEC C V6.0-001 on OpenVMS> Alpha V7.1).  I have tried to check out the sites mentioned in; the FAQ (www.sidhe.org, genetics.upenn.edu) but can't reachS: either site (timeouts).  Any help or pointers appreciated.   thanks      L * Sent from RemarQ http://www.remarq.com The Internet's Discussion Network *G The fastest and easiest way to search and participate in Usenet - Free!D   ------------------------------  # Date: Fri, 28 Apr 2000 17:24:59 GMTh3 From: Eric Dittman <dittman@narnia.int.dittman.net> % Subject: DCPS and TI Microlaser Pro Ex@ Message-ID: <LPjO4.14987$Nd6.2535499@news-west.usenetserver.com>  = Does anyone have a TI (now owned by Genicom) Microlaser Pro E.: working with DCPS?  I have the printer connected through a> DECserver 700 and can telnet to the DECserver port and control? the printer that way, but when I try to use DCPS the data light ? on the printer flashes and the LCD display shows processing butc? nothing happens.  Thinking that at 9600 baud I may have to wait6> a while I left it going for a couple of hours.  When I checked back it was still processing.t   I have a DCPS-OPEN license.$ -- @ Eric Dittman dittman@dittman.nete   ------------------------------  # Date: Fri, 28 Apr 2000 17:20:06 GMTw- From: DRHarrold@Earthlink.net (David Harrold)t) Subject: Disk I/O Statistics Don't Add Upa9 Message-ID: <8F247988EDRHarroldearthlinkne@207.217.77.21>n  E We are looking at some of the data captured from Monitor and graphinga the results for review.   E One of the things we have noticed is a couple of shadow sets report aeE certain number of I/O's, but the member volumes I/O count doesn't addcC up to that number.  We are expecting them to add up and most of the D disks on the system do.  Specifically, we're looking at the average  numbers.  E For this example, I've picked the system disk, DSA0:.  The shadow sethF is made up of 2 disks served off of separate HSJ50 controllers.  This E data is being captured off of a GS60E running VMS V7.1-2.  This is a 5  cluster of 3 8400s and 2 GS60Es.  3                             OpenVMS Monitor Utility 1                               DISK I/O STATISTICSSP                                  on node HNAC         From: 27-APR-2000 00:00:11P                                     SUMMARY           To:   27-APR-2000 00:15:11  O I/O Operation Rate                         CUR        AVE        MIN        MAXA  O DSA0:                  AXPVMS71         104.06      86.87      34.06     127.80n  3                             OpenVMS Monitor Utilityr1                               DISK I/O STATISTICS P                                  on node HNAC         From: 27-APR-2000 00:00:11P                                     SUMMARY           To:   27-APR-2000 00:15:11  O I/O Operation Rate                         CUR        AVE        MIN        MAXt   [...]aO $1$DUA0:      (EEYORE) AXPVMS71           3.06       0.52       0.00      15.66e [...]s  3                             OpenVMS Monitor Utilityl1                               DISK I/O STATISTICSoP                                  on node HNAC         From: 27-APR-2000 00:00:11P                                     SUMMARY           To:   27-APR-2000 00:15:11  O I/O Operation Rate                         CUR        AVE        MIN        MAXe [...] O $1$DUA24:     (HUEY)   AXPVMS71           3.06       0.51       0.00      15.66t     So the questions are:S  J 1.)  Is it right to assume that these numbers should add up?  If not, why & do most of the other disks not add up?  6 2.)  If they are supposed to add up, why are they not?     Thanks,s   Dave Harrold   ------------------------------  % Date: Fri, 28 Apr 2000 15:47:25 +0930)% From: Jeremy Begg <jeremy@vsm.com.au> . Subject: Re: Disk subsystems for OpenVMS Alpha* Message-ID: <39092CF5.214C0951@vsm.com.au>   Hello Hoff,    Hoff Hoffman wrote:: > T > In article <3906477D.12EC9A09@vsm.com.au>, Jeremy Begg <jeremy@vsm.com.au> writes:I > :I am working with one of my customers to upgrade their VAX system to aeN > :suitable Alpha configuration.  Having read the hardware literature I have a, > :couple of questions about configurations. > :vN > :The existing system is a VAX 4100-A with 128MB RAM and approximately 9GB ofO > :disk, with a TZ877 for backup.  The disk is organised into shadowed volumes: M > :3 pairs of RF36s plus one pair of RZ29s.  The system disk is an additionalg% > :shadow set, being a pair of RZ26s.t > B >   The DSSI interconnect is roughly equivilent to SCSI-1 in termsG >   of performance.  I/O performance is *significantly* better with thee, >   UltraSCSI (Ultra1 and Ultra2) widgets...  < I didn't realise the DSSI bus & drives were quite that slow!  s  M > :The software is "TOLAS", a warehouse management application which uses RMS N > :Prolog-3 Indexed files for the database, and connects to another system via- > :IBM"s MQ-Series message queueing software.o > E >   You'll want to see that TOLAS and MQ-Series are available in youre! >   target porting environment...m  I Yes, TOLAS, MQ-Series and assorted other components are all available for L OpenVMS Alpha.  There is one piece, a report writer, which could be in doubtK but we know of a replacement which offers far greater functionality anyway.m    H >   As much as I'd prefer to see this customer move to an AlphaServer DSI >   series system, I'd first look at the RMS file structure and tuning it J >   and optimizing the RMS files.  I'd then look at a VAX upgrade to a VAXI >   4000 model 108 -- the last order date for this system is in September-4 >   of 2000 -- rather than an OpenVMS Alpha upgrade.  K It's mainly disk I/O & CPU which are the limiting resources.  The 128MB RAMaG is also well-used, mostly for VIOC and global buffers.  The RMS indexed K files are purged (i.e. old records deleted) and CONVERTed on a weekly basis K and the FDL descriptions which control the CONVERT operations were reviewedu
 not long ago.k  G I agree that an upgrade to a new VAX would be simpler, but if it's onlyiI twice as fast it's not going to be a big improvement on what they've got,gJ given that the current system is running out of CPU power and they want to  significantly increase the load.  I Also the SCSI options for the 4108 seem to based on DSSI via the HSD10 so K I wouldn't think there was much improvement in I/O throughput if DSSI is as K bad as you say.  I suppose we could go for the new "VAX Adapter" to hook upaD UltraSCSI but if the system turned into a cluster (which is a remoteI possibility) we'd be back to using DSSI for shared storage, would we not?     G >   I will also assume that you are aware that one of the prerequisites F >   for OpenVMS Alpha is sufficient physical memory, and that you haveH >   sized the 1GB based on current use and growth expectations.  OpenVMSH >   can and does prefer to use physical memory for caching and for otherE >   performance-related operations, and OpenVMS Alpha inherently uses ! >   more memory than OpenVMS VAX.A  I It's funny how a virtual memory system performs better with more physicalC
 memory :-)  K The present system has 128MB RAM which is just enough for the current load. F I'm stating 1GB on the basis that Alpha will use 3-4 times as much RAME (compared to the VAX) as a matter of course, plus allowing a generous 4 overhead for extra caching and buffering and growth.    J >   A worry over power supplies is a little odd -- if that sort of concernG >   is involved, I would typically look at using an OpenVMS Cluster andA >   multiple hosts...   F Perhaps I'll have another look at the price of cluster licences.  FromG memory I thought that they would not be cost-effective but I'll have to 
 confirm that.A  / > :In theory a DS20 would do just as well but I I > :figure if they're going to buy a new system they might as well buy theO > :latest hardware!  > G >   You can hang quite a bit of storage off the AlphaServer DS10 series ? >   box, what with the HSZ and HSG series controller widgets...   K But isn't the DS10 limited for I/O?  For example, if the system does evolve = into a SCSI cluster, I'd prefer to use two SCSI buses for theFH disk subsystem (for performance & reliability) which on a DS10 leaves noE SCSI bus available to the tape drive (the DS10 supports only two SCSI : controllers).  And the DS10L has only one PCI slot anyway.  N > :I'll need one SCSI bus for the TZ877 (I see no reason why we can't continueL > :to use it for backup) but would prefer not to have the data disks on that > :bus.U > L >   The TZ89 series is a better and faster DLT widget -- as you go to largerK >   and larger disks, most folks prefer not to see a corresponding increase K >   in the BACKUP processing window.  When looking at the tapes, also check : >   the sustained device throughputs to see what I mean...  K Actually downtime for backup isn't a big issue with these people so we'd beIJ prepared to try it out with the existing drive and replace it only if it'sK absolutely necessary.  It's a pity that the "Snapshot Services for OpenVMS"03 got scrapped, as it would have done the job nicely!s  E > : To ensure acceptable performance I think a RAID solution would be H > :best, so I have to decide between an embedded RAID controller such asK > :the KZPAC (aka Raid Array 230/Plus) or an external subsystem such as theDG > :RA3000.  In addition, for maximum data redundancy, the RAID would berE > :configured as striped shadowsets (or is it the other way around?).s > B >   Be aware that most (all?) backplane RAID controllers generallyD >   cannot be configured into a multi-host SCSI environment, if that >   becomes of interest...  J Yes, I was aware of that.  I've also since learned that the backplane RAIDJ controllers place upper limits on the amount of data than can be served upK to the host (max 32GB) which is not much above my proposed configuration of K 27GB.  So the RA3000 approach is looking better and better.  Incidentally I7J arrived at a figure of 27GB based on three 9GB drives.  I realise it's notE an awful lot of space these days but it is more than three times whatoF they're currently using so it should be sufficient.  At least with the) RA3000 we can easily add more if need be.u  D >   If you do configure a multi-host SCSI environment, you will alsoF >   want to look at the cluster communications interconnect -- be that: >   Fast Ethernet, Gigabit Ethernet, FDDI, or otherwise...  I I suspect that the 100MB/sec ethernet controllers now available should be J fast enough for this application.  I'll also check out the price of memory channel as an alternative.  d  I >   The HSZ70 and HSZ80 controllers used in various recent storage arraysaC >   can span SCSI buses within the same OpenVMS Cluster, assuming aeH >   sufficiently recent OpenVMS release and current controller firmware.  H What about the HSZ22, as used in the RA3000?  We will be running OpenVMS
 Alpha V7.2-1.d   Thanks,r   	Jeremy Begg   ------------------------------  % Date: Fri, 28 Apr 2000 02:56:04 -0400y0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>. Subject: Re: Disk subsystems for OpenVMS Alpha/ Message-ID: <390935FF.42102E40@vl.videotron.ca>a   Jeremy Begg wrote:D > >   The DSSI interconnect is roughly equivilent to SCSI-1 in termsI > >   of performance.  I/O performance is *significantly* better with thee. > >   UltraSCSI (Ultra1 and Ultra2) widgets... > > > I didn't realise the DSSI bus & drives were quite that slow!  M Wasn't DSSI an almost exact replica of SCSI ? (But different enough that onlyr DEC drives would plug in ?)   ( Also, while on the topic of disk drives.  K I have a DILOG SQ-739 (I think) SCSI-Q-BUS controller. It makes disk driveshT appear as RA to VMS. A 250meg drive appears as a RA80 and a 675meg drive as an RA82.  L Question: is there an architectural limit to RA drives that would prevent meM putting in a 2 gig drive ? I realise that a Vaxstation 3100 has problems withy( a system drive being bigger than 1 gig.   L Would a Microvax II outfitted with a Q-bus SCSI controller that makes drives6 appear as RA have problems booting off a large drive ?   ------------------------------  # Date: Fri, 28 Apr 2000 12:13:03 GMTt( From: Terry Kennedy <terry@gate.tmk.com>. Subject: Re: Disk subsystems for OpenVMS Alpha' Message-ID: <Ftq79r.EJF@spcuna.spc.edu>p  ' Jeremy Begg <jeremy@vsm.com.au> writes:iE > The system I am going to propose to the customer will consist of an M > AlphaServer DS20E with 1GB RAM and 27GB disk.  (I am recommending the DS20EeN > rather than a DS10 or DS10L because of the extra I/O capabilities offered byI > more PCI slots and two PCI controllers instead of 1, and because it haseI > multiple power supplies.)  In theory a DS20 would do just as well but IiH > figure if they're going to buy a new system they might as well buy the > latest hardware!  H   The DS20 series gets you a faster memory subsystem as well as more PCI slots compared to the DS10.8  M > I'll need one SCSI bus for the TZ877 (I see no reason why we can't continue K > to use it for backup) but would prefer not to have the data disks on thatoI > bus.  To ensure acceptable performance I think a RAID solution would be G > best, so I have to decide between an embedded RAID controller such asdJ > the KZPAC (aka Raid Array 230/Plus) or an external subsystem such as theF > RA3000.  In addition, for maximum data redundancy, the RAID would beD > configured as striped shadowsets (or is it the other way around?).  @   I have 2 DS20's with the KZPAC and they work fine (see below).  H > Originally I had thought that I would use the KZPAC-CA variant (3-portL > UltraSCSI RAID) with one port connected to the DS20E's internal disks, andM > the two other ports connected to a split-bus BA356 StorageWorks shelf.  ButtM > my reading of the DS20E QuickSpecs tells me that the internal disks are not J > supported with the KZPAC and the KZPAC won't work with a split-bus BA356K > anyway.  So that means either using a KZPAC-AA with all six data disks oniL > one SCSI bus, or a KZPAC-CA with two BA356s, each BA356 holding three dataI > disks, and the third port on the KZPAC being unused.  The latter option 9 > seems to make the most sense in terms of I/O bandwidth.   H   The documentation is very, very unclear on this. On the DS20 (not sureF about the DS20E), you can indeed run the internal shelf off the KZPAC.I And you can use the bus split kit (labeled as "AS1200 bus split kit", but=F it works fine on the DS20) to bring 2 of the card's 3 channels to the K internal shelf. Of the 2 systems I ordered in this configuration, the first.I one (October 1999) came with the bus split kit "on the side" and I had toxK install it. The second system (February, 2000) came with it factory-instal-= led.  I   I have my 2 DS20's with 7 9GB drives in each, 4 on one bus and 4 on thel> other. One of them also has 6 4GB drives in an external BA3xx.  I   You'll still need a single-ended SCSI adapter (usually the Q-logic one)4J to support the internal CD-ROM (the KZPAC "notices" CD-ROMs but won't makeG them available to the system), so you can hang your TZ tape off that asaG well. There's a bogus restriction on mixing internal and external SCSI  J devices on the same controller, but it works fine (and the controller man-I ufacturers don't know why Digipaq is saying this). I suspect it's because 8 it makes it easier to goof when setting the termination.  I   A later note shows you've been confused by the literature on the KZPAC.fG The total amount of disk storage it can serve to the host is 256GB, butiK it can have over twice that connected - it can serve up to 8 logical drives,K (which appear as DRAn:) and each of those logical drives can be up to 32GB.oK Extra disks in RAID 1 mirror sets don't count, nor do hot spare drives. TheiI limit is entirely in the amount of data served to the host. I have mostlyeK RAID 5 sets, and one controller has 87GB attached - 7 9GB drives internallye= on 2 of the buses and 6 4GB drives externally on the 3rd bus.P  - 	Terry Kennedy             http://www.tmk.comP5         terry@tmk.com             Jersey City, NJ USA0   ------------------------------   Date: 28 Apr 2000 16:32:18 GMT! From: briannfo@aol.com (BrianNFO)h) Subject: DLT Drive Software Write Lockingn: Message-ID: <20000428123218.14595.00000548@ng-ch1.aol.com>  L I have a pair of TZ887's connected to an HSJ52.  All of a sudden, the drivesO are write locking whenever you load a tape.  From a hardware standpoint, thingsiK look fine.  The tapes are NOT write protected, and when the tape loads, the J drive "write protect" light is NOT on.  When you mount the tape to VMS, itN indicates the device is write locked...and a show dev/full indicates the driveO is software write locked.  (This happens on multiple tapes...both brand new and0 used.)  J This is happening on 2, TZ887's that before now we're perfectly fine.  Any/ thoughts on what is causing this?  Thanks much.7   Brian4  
 P.S.  VMS 6.2e   ------------------------------  , Date: Fri, 28 Apr 2000 15:55:29 +0200 (CEST): From: "Gotfryd Smolik, VMS lists" <gotfryd@stanpol.com.pl>@ Subject: Re: Does backup command backup used or allocated space?J Message-ID: <Pine.LNX.4.10.10004281548530.1437-100000@irys.stanpol.com.pl>  ' On Tue, 25 Apr 2000, Tom Steuver wrote:n  M +When using the backup command in OpenVMS 7.2-1, does the utility take up the M +file's used block space or allocated block space on the tape?  If a disk hass9 +a higher cluster size, does it tape more tape to backup?     Why not make a short test ? $ h/out=t.t *... $ set file/att=ebk=1 t.t $ dir/siz=all t.t    Directory DISK$USER:[GS]  $ T.T;1                      1/17205     $ back t.t x.x/save/logs, %BACKUP-S-COPIED, copied DISK$USER:[GS]T.T;1 $ dir/siz=all %e   Directory DISK$USER:[GS]  $ T.T;1                      1/17205  $ X.X;1                    126/126       ...and you have the answer.l6  While restore the "free place" is allocated, BUT only7 filled with unusefull info (random or zeroed, regardingm the highwater volume setup).9  How the size is rounded (when the cluster size on backeda9 up and restored disk are differrent) you can control witha /TRUNCATE qualifier.   +Thanks, +Tom Steuver    Regards - Gotfryd   --E =====================================================================aF $ ON F$ERROR("LANGUAGE","ENGLISH","IN_MESSAGE").GT.F$ERROR("NORMAL") - 		THEN EXCUSE/OBJECT=ME . $!                        GS@stanpol.zabrze.plE =====================================================================    ------------------------------  # Date: Fri, 28 Apr 2000 09:29:30 GMT 0 From: carlini@true.lkg.dec.com (Antonio Carlini)* Subject: Re: Dropping DECnet..don't do it!* Message-ID: <8ebl2v$bt3@usenet.pa.dec.com>  a In article <38FF3714.8B567F7F@vl.videotron.ca>, JF Mezei <jfmezei.spamnot@vl.videotron.ca> wrote: N >NCL could then simply become an emulator that would translate NCL commands toN >NCP commands to provide compatibility with those already running NCL scripts.  M Phase V uses CMIP to manage stuff so although you could improve the existing  H NCP emulator somewhat you cannot actually manage Phase V using Phase IV N concepts - the two don't match. You might as well ask for an NCP interface to  manage your IP stack.5  I >My impression of decnet-5 is that it lacks a management tool that was asoJ >simple and easy to use as NCP. I got rid of decnet-5 and upgraded back toJ >decnet-4 because I was sick and tired or trying to fight it to remove all. >those opcom messages. It just wasn't natural.  N I believe I've stated before that I've heard that users were originally never L intended to see raw NCL as such. They were supposed to use some sort of GUI O interface (possibly as part of DECmcc). I don't know where along the line that t. idea was dropped (assuming it was ever there).  L My first exposure to NCL was doing some of the SNMP work on DECnis. At that L stage I could use NCP in my sleep and NCL certainly seemed new and strange. M But it was no different to having to sit down and understand how to drive an gN IP stack. I doubt that NCL itself is the problem (other than being different) H - I suspect its more to do with the fact that NCL exposes the different N protocol layers more explicitly. This is bad in that you need to look at what M is happening in a different way but is good in that you can generally follow m3 what is going on all the way up and down the stack.    Antonioe  I Antonio Carlini                            Mail: carlini@true.lkg.dec.comT# DECnet-Plus for OpenVMS Engineeringt6 COMPAQ                                     Reading, UK   ------------------------------  % Date: Fri, 28 Apr 2000 09:18:23 +0200o= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>a- Subject: Re: F$File_Attribute from a program? ) Message-ID: <39093B3E.E89C8FFB@gtech.com>l   "Richard L. Dyson" wrote:fH >         I would like to find the creation date of a file from within aH > C program on an OpenVMS system.  For DCL, I can use a lexical such as: > - > x = F$File_Attributes ("<filename>", "CDT")  > J >         Can someone suggest a system service or run-time library routineH > to accomplish the same thing?  Or a C routine that I have missed would > be > OK too. :)  7 C RTL stat or RMS XAB's (XABDAT if I remember correct).    Arne   ------------------------------  % Date: Fri, 28 Apr 2000 09:20:00 +0200 = From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com> - Subject: Re: F$File_Attribute from a program?i) Message-ID: <39093BA0.72201AB8@gtech.com>f   Terry Marosites wrote:K > Try lib$get_command to call the = F$File_Attributes ("<filename>", "CDT")i   ????  = The poster wanted code to do it, not a spawn hack to use DCL.t  8 And BTW: LIB$GET_COMMAND is irrelevant in this context !   Arne   ------------------------------  % Date: Fri, 28 Apr 2000 08:37:02 -0400a[ From: jamese@beast.dtsw.army.mil (Ed James, TCS Inc, 410-295-1919, ed.james@telecomsys.com)p- Subject: Re: F$File_Attribute from a program?t0 Message-ID: <00042808370237@beast.dtsw.army.mil>  H "Thomas nilsen" <Thomas.Nilsen@kverneland.com> wrote On Fri, 28 Apr 2000< 08:58:44 +0100, in <XAbO4.13218$By1.239098@news1.online.no>:  M > Is there any way to set up a print queue on OpenVMS AXP v7.1 so that prints N > sent to this queue is saved as file on the system instead of going to a real
 > printer?  H Then Christoph Gartmann and Arne Vajhoej suggested writing or adapting a
 symbiont.   H Well, at the same time, David B Sneddon <dbsneddon@bigpond.com> wrote inA a reply to a different thread about his library available at URL:bE http://www.users.bigpond.com/dbsneddon/software.htm, which contains amE null device symbiont. Set the spool directory where ever you want andl/ don't specify /DELETE, and there you are. YMMV.S   Ed    : Ed James                           ed.james@telecomsys.com5 TeleCommunications Systems, Inc.   voice 410-295-1919o5 275 West Street, Suite 400         fax   410-280-1094w Annapolis, MD 21401-1740   ------------------------------  % Date: Fri, 28 Apr 2000 09:58:04 -0400 % From: Bob Harris <harris@zk3.dec.com>t- Subject: Re: F$File_Attribute from a program? 3 Message-ID: <280420000958045623%harris@zk3.dec.com>d  / [[ This message was both posted and mailed: see ;    the "To," "Cc," and "Newsgroups" headers for details. ]]V  C In article <8e9sj4$1fb$1@mailint03.im.hou.compaq.com>, Hoff Hoffmane& <hoffman@xdelta.zko.dec.nospam> wrote:  = >In article <39082A2E.682F79AB@uiowa.edu>, "Richard L. Dyson"' ><rick-dyson@uiowa.edu> writes:lG >:       I would like to find the creation date of a file from within ac# >:C program on an OpenVMS system...rI >:       Can someone suggest a system service or run-time library routinerH >:to accomplish the same thing?  Or a C routine that I have missed would >:be OK too. :)o >tE >  There are various ways, but the C stat and fstat calls are likely hH >  among the easiest for you in this particular application environment.F >  Look in the field st_ctime in the returned stat_t data structure...  B I would never want to argue with Hoff, but stat.st_ctime on a UNIXE system is not the create time.  The POSIX standard says that st_ctimeaF is "time of last status change".  On UNIX this is essentially the last< time some attribute of the file was changed, for example its! protections, or its owner, etc...   > And here is where Hoff has the advantage on me, OpenVMS may do8 something that is non-standard with the st_ctime field.   A I just wanted to question the correctness of the use of st_ctime.a   ------------------------------   Date: 28 Apr 2000 15:18:10 GMT2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)- Subject: Re: F$File_Attribute from a program? 6 Message-ID: <8eca3i$3cv$1@mailint03.im.hou.compaq.com>  [ In article <280420000958045623%harris@zk3.dec.com>, Bob Harris <harris@zk3.dec.com> writes:hC :I would never want to argue with Hoff, but stat.st_ctime on a UNIXrF :system is not the create time.  The POSIX standard says that st_ctimeG :is "time of last status change".  On UNIX this is essentially the lasta= :time some attribute of the file was changed, for example itsm" :protections, or its owner, etc...  @   On OpenVMS, there exists ctime, mtime, and atime, for the fileB   creation, modification, and last access time.  This is obviously   different from POSIX.y  A   I know OpenVMS doesn't (yet) have the last access time support, (   that's something we're looking to add.  C   If you're operating in a POSIX-compatible environment on OpenVMS, ;   then the current ctime behaviour clearly isn't for you...   B :I just wanted to question the correctness of the use of st_ctime.  C   I don't have an ANSI C standard immediately visible in the stackswD   in my office to check what _that_ says about this -- the base RTL ;   behaviour is targeting ANSI C, and not currently POSIX...c    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Fri, 28 Apr 2000 15:59:43 +0300m From: pkassapi@panafon.grr* Subject: File Transfer Management software6 Message-ID: <422568CF.0048E7D2.00@mailsmtp.panafon.gr>  
 Hello all!  P The company I work for is about to install a new Wholesale Billing software on aF couple of AlphaServer GS140 systems running OpenVMS Alpha V7.1-2. ThisK application is expected to produce over 400 files on a daily basis. Most ofdM these files ( > 200) are to be transferred via FTAM over X.25 lines to remote N systems for processing. The rest of them will be transferred to local systems.J We are looking for a software tool that can be used to automate these fileN transfers. Ideally, it would sit on top of DECnet-Plus and UCX and using COPY,B COPY/APPL=FTAM and FTP (or their respective APIs) will be able to:  M - manage multiple concurrent outgoing (maybe incoming too) file transfers via & FTAM, FTP or DECnet over WAN/LAN linesK - manage remote user details (usernames, passwords, remote network address,s etc.)nH - automatically/manually retry failed transfers up to an arbitrary (user configurable) number of timestN - write these files to ISO 1001 formatted reel tapes for dispatching to remote parties upon operator requestg= - retransmit successfully transferred files upon user request:K - monitor file transfer activities and provide necessary logging facilitiesDF - be easily extensible to include additional file transfers methods to0 additional destinations whenever the need arisesJ - produce reports of successful/failed file transfers in as much detail as possibleK - provide different levels of user access to the file transfers it controls2& (user, operator, administrator, etc..)  G Any additional features, are welcome of course. Do you know of any suchlI software? Any pointers you can provide to its whereabouts will be greatly  appreciated.   Thank you in advance,i
 Panayiotis   ---------------- Panayiotis Kassapidisr Systems Engineer+ IT/Computer Systems Centre/Systems Servicesv+ PANAFON Hellenic Telecommunications CompanyR e-mail: pkassapi@panafon.gr    ------------------------------  % Date: Fri, 28 Apr 2000 10:46:13 +0100d  From: steven.reece@quintiles.comI Subject: Re: Including CD media with systems? (was Re: OpenVMS marketing) > Message-ID: <802568CF.0035EADA.00@qedilc01.qedi.quintiles.com>  P I can understand the rasons for not wanting to ship media with every system thatP goes out - I can well remember the problems caused by copies of DOS 3.3 on 3 andO a half inch floppy disks shipping with every system at college.  It made a nice.G pile of reusable disks, provided they still had their write enable tabsy	 in.......o  ? My main problem with sales materials is a little unrelated.....S  L When looking for technical sales information I find that I still rely upon aH _paper_ copy of the Systems and Options Guide.  The one I've got here isJ November 1997 so it's a little long in the tooth these days but it's still quicker than looking elsewhere.2  N What I'd like is the restarting of the SoG/SoC or something else that providesK me with the same level of information.  When one has big Alphas around it'ssO often not a problem of things changing so quickly that before the book has beennM released to customers/VARs it's out of date, unlike the PC world that changeseJ rapidly due to marketing pressures and the need for speed to run MicrosoftP bloatware.  It makes it easier for the VARs as well since I can call them up andI ask for prices and availability of option AB-CDEFG-HJ which is a whateveraM widget.  I get my quote quicker which means that Compaq get the order quickeriN (provided the boss agrees to the purchase of course) which helps the cash flow
 at Compaq.  N It also saves me having to trawl round a web-site in which it can be difficultO to find the particular part number or description of the widget one requires oraO the page which tells you that this is a legal configuration.  This latter pointkP comes from the (apparent) fact that one web-page doesn't suit everyone - someoneO who is non-technical may be able to find what they want really quickly from theoO Compaq web pages whilst I am usually after techie specs and such like which can P be difficult to find on the same pages.  (I'd also add that this is not a CompaqP specific issue - I have problems getting the nuts and bolts infomation I require from _many_ vendor sites).   Steve.   Hoff wrote:tK >>>  The "solution"?  If you need any particular CD media with your system,hH   please order it.  We do try to make this prerequisite for at least oneI   media kit on-site clear in the sales materials provided -- if it is not G   clear, please let me know and I will pass the comment(s) along to ther&   appropriate product manager(s)...<<<   ------------------------------  # Date: Fri, 28 Apr 2000 14:57:48 GMTh% From: Alan Greig <agreig@my-deja.com>bI Subject: Re: Including CD media with systems? (was Re: OpenVMS marketing)t) Message-ID: <8ec8t1$vnt$1@nnrp1.deja.com>i  > In article <802568CF.0035EADA.00@qedilc01.qedi.quintiles.com>,#   steven.reece@quintiles.com wrote:hG > When looking for technical sales information I find that I still relyo upon aG > _paper_ copy of the Systems and Options Guide.  The one I've got heret isF > November 1997 so it's a little long in the tooth these days but it's still ! > quicker than looking elsewhere.Z  B Compaq/DEC did continue to produce the paper SOC for a while afterA announcing they wouldn't. I have a 1999 SOC in the standard bounds@ format which includes systems such as the DS20, ES40 and XP1000.  E Worth tracking a copy down but I think that's the last one though :-(    --
 Alan Greig    & Sent via Deja.com http://www.deja.com/ Before you buy.t   ------------------------------    Date: 28 Apr 2000 10:17:47 -0400+ From: randall.burlew@srs.gov (Randy Burlew)t' Subject: Re: Ingres and Defragmentation8, Message-ID: <2000Apr28.101747.12360@srs.gov>   Bob Kaplow says... >v> >In article <38F6465E.6EB60E3B@home.com>, David Warren writes:= >> We have used Diskeeper by Executive Software for somewherea9 >> around 6 years with Ingres.  We aren't doing real-timet< >> data aquisition, though.   Mileage may vary, just another >> pointer.  > ? >Diskeeper says that their defragmenter works fine with INGRES.e > ; >CA says do not run a defragmenter on INGRES disks. I don't ; >think you could find a product that is worse on a VMS file ( >system than INGRES. Especially pre-7.2. >o: >We run Diskeeper on our production system, but not on the< >INGRES disks. IIRC we've run it against INGRES disks on our; >development system without any incidents. Virtually all of : >our production system crashes are due to INGRES bugs, not" >Diskeeper, and certainly not VMS. >g >YMMV. >n >        Bob Kaplow         ; We have used Diskeeper on our Ingres production systems for 9 years with no problems. I have, in the past, specificallya8 asked technical support about this and have been assured there is no problem.  7 I would be interested in knowing what you were told and : the problems that could result from running a defragmenter on INGRES disks.   Randy Burlew randall.burlew@srs.gov* ******************************************* Anything I post is my opinion and does not* represent the views of management. They do not like Ingres or VMS.d* ******************************************   ------------------------------  % Date: Fri, 28 Apr 2000 12:05:20 +0100-3 From: "Jefferson Humber" <matrix01@globalnet.co.uk>aK Subject: Re: Is it possible to get rid of this annoying boot-up message ???a# Message-ID: <8ebr9f$rb6$1@gxsn.com>h   Thanks,a   That has down the business.s   Cheers,o   Jeff= "JF Mezei" <jfmezei.spamnot@vl.videotron.ca> wrote in message ) news:3907A35A.5B94E243@vl.videotron.ca...n > Jefferson Humber wrote:/0 > > > Check and see if WINDOW_SYSTEM is still 0.3 > > > I have seen the DECwindows Startup change it.  > L > Have you got a $DEFINE DECW$IGNORE_DECWINDOWS TRUE  in SYSTARTUP_VMS.COM ? If8 > not, it may still try to start the windowing software.   ------------------------------  % Date: Fri, 28 Apr 2000 15:19:36 +0100 - From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>eH Subject: Re: Multiport serial card on PC as AlphaServer serial consoles?) Message-ID: <39099DF8.B15D1B20@bbc.co.uk>u   "Dale A. Dellutri" wrote:r   >d >h: > Any comments?  Anybody doing this?  Issues?  Criticisms?  ? I would consider security an issue worth considering deeply, ifTD you plan to use your network connected desktop PC for this. Just who7 exactly might be able to connect to your console ports?A   >E >E > --( > Dale Dellutri -- ddellutr@enteract.com   --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.u   ------------------------------  % Date: Fri, 28 Apr 2000 12:52:20 -0400.5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com>c  Subject: Re: OpenML for OpenVMS?+ Message-ID: <8ecfgo$5bt$1@lead.zk3.dec.com>b  K There are really two 3D standards right now.  OpenGL, which has been aroundeK a while and is widely used in professional 3D graphics - and is the UNIX 3DoF standard of the day.  And Direct3D, which is the Microsoft Windows NIH* standard that they want to replace OpenGL.  J If OpenML, or Bob's me-too-3D becomes a widely accepted standard, and moreL importantly - it's at least "compatable" with an X11 implementation - then IB would expect at some time that you would see a VMS implementation.  K But old standards die hard.  VMS is still selling and supporting PHIGS, thek- last great 3D standard before OpenGL arrived.       F Rudolf Wingert wrote in message <200004271439.QAA06932@fom.fgan.de>... >Hello,e >aJ >I have heard, that 3dfx, ATI, 3DLabs, Intel, IBM, Compaq SGI and S3 wouldK >like to create a new multiplatform standard competing to DirectX. Is there)B >any chance, that there will be an OpenML for OpenVMS like OpenGL? >i >TIA and regards Rudolf Wingertp >r   ------------------------------  % Date: Fri, 28 Apr 2000 16:07:02 +0200 $ From: Paul Sture <sture.ch@sture.ch>4 Subject: Re: OpenVMS and WNT integration for Dummies+ Message-ID: <VA.0000001f.16b29d5c@sture.ch>   B In article <795557946wnr@natron.demon.co.uk>, Roger Barnett wrote:0 > From: Roger@natron.demon.co.uk (Roger Barnett) > Newsgroups: comp.os.vms 6 > Subject: Re: OpenVMS and WNT integration for Dummies' > Date: Thu, 27 Apr 2000 12:06:52 +0100p > U > In article: <VA.0000001b.0dc21acd@sture.ch>  Paul Sture <sture.ch@sture.ch> writes:a > >  > > # > > .. after rewriting with COM ...  > > ? > > "Here's the best part: Each order taker get his or her own oH > > instantiation of the COM object, so each customer request locks the ; > > database - that is, reserves each item - in real time."a	 > >[snip]  >  > D > No doubt the book provides equal coverage of alternative solutions" > such as using CORBA and/or EJB ? > Q They do get a brief mention, but that's about it. COM gets the main mention, but "= Reliable Transaction Router also gets a good amount of space.   F > Or is it implicit in the "Dummies" that only Microsoft technology is > considered ? > S Now I've thought a bit more about it, it's probably not such a bad idea to fill it  O with M$ terminology - talk to those who've only ever heard of M$ products in a  V language they understand, but _do_ introduce them to concepts such as RTR at the same  time.I  
 ..next day...)  S Well instead of just dipping into the book I started reading it from the beginning  O last night, and I have changed my mind. There is loads of meat in there and it lR manages to introduce features we all know and love in a way which should grab the P attention of those experienced in NT without insulting their beloved OS head on.  U Yes, nicely done. I _will_ post a review of it once I've finished reading it, and at nH the moment I think it's going to be a lot more positive than I expected. ___-
 Paul Sture Switzerland    ------------------------------  % Date: Fri, 28 Apr 2000 12:57:33 -0400s. From: Michael Austin <maustin@nc.prestige.net>4 Subject: Re: OpenVMS and WNT integration for Dummies- Message-ID: <3909C2FD.D98A87@nc.prestige.net>   , This is a multi-part message in MIME format.& --------------0868DF7BC2855FD31DD494F1* Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit   H I am a DBA (Rdb and Oracle/Oracle if I must) and if the code was writtenF properly, first come/served is always possible -- Even using COM.  YouJ still have the same problem. What planet is this guy from?  Obviously thisK book was written to give some unwitting bean-counters the idea that you canrI do the same thing in WNT that you can in OpenVMS.  Hogwash.  I still have H not seen ANY WNT server that could sustain the TPS-anything numbers thatG you can on any OpenVMS system.  And to say that VMS is legacy is also a'F bunch of hog slop.  OpenVMS will do as much or more today than ANY WNTF server and do it without crashing or having to reboot every other day.K Just because it has a GUI interface does NOT imply that it is better.   And1H if you want GUI, web-enabling your apps can be done fairly easily if you= know what you are doing.   And guess what?  It doesn't crash.d  I Hmmm,  I was thinking,  look at the enormous cost of deploying a WNT app,0K such as each PC must have a bunch of code running on the desktop to make itdA all work.  So in looking at "cost savings" they are investigatingVI "Thin-Client".  And you can't get a thinner client than a VTxxx terminal.h :)  I All of that being said,  it does not sound like they are trying to do  as,6 much integration as replacement. (more comments below)   Michael Austin Michael Austin, Inc. A DBA Consulting Firmw 704-947-1089 (Office)n 704-236-4377 (Mobile - 24x7) michaelaustininc@hotmail.com     >c9 > Too US American for me, hey Puh-Leeze wait for me, Huh?i > G > blah blah - successful company gets too many orders which they cannotaC > process in order because they "used terminal emulators to fill ong? > on-screen character-cell forms for their legacy application."  >L > The next key words are:n >eD > "But the order takers found that they were not seeing when catalogC > items were out of stock because the application handled the orderh5 > transactions on a "first come, first served" basis.s >u  H It is called read consistency.  You must look at the data as it was whenA you started your transaction.  Reading "dirty data" (data not yet.J committed), although possible with some database engines is not preferableK from a pure data standpoint.  What if that user does not commit, but electsoK to abort the transaction. Now you have told a customer you are out of stock B when you actually have it.   Airline reservation systems allow theC reservationist to reserve more seats than are available on a plane.sD Catalog companies also allow products to be over-sold and it is then: back-ordered.  Event ticket sales do not allow seats to beH "double-booked".  The record for that seat will be locked when the FIRSTA reservationist request the seat (exclusive lock) so that no otheraH reservationist can attempt to sell the seat while the customer is in theJ process of selling a ticket for that seat.  The next resvationist attemptsJ to lock the same record, will be given a message that it is not available.  G All of this has NOTHING to do with whether or not they are on a GUI/COM J application on a character-cell-based application.  It has EVERYTHING withJ how the application and database does record-level locking.  And no matterJ what they say... it is always on a first-come/first-served basis.  If not,' then you have some serious data issues.r   >r! > .. after rewriting with COM ...t >t< > "Here's the best part: Each order taker get his or her ownE > instantiation of the COM object, so each customer request locks thef9 > database - that is, reserves each item - in real time."  >e  E The first one to commit the record will still win.  Everyoneelse will-C either be back-ordered or denied updates.  first come/first served.   6 > Sounds like they were rewriting a CICS system to me. >aG > Sorry, but my team and I implemented a first come, first served order1D > entry system back in 1977-8. With record locking. (RT-11 & Dibol -B > CTS300, if you must know). We later rejected IBM's CICS proposalI > because after saying to a customer "that item is yours", their solutioneG > was that by the time you hit <enter> (oops sorry, EOB - End Of Block)p > it could be somebody else's. >oI > Oh, and another thing, (and as someone raised to speak and write what IiH > prefer to call British English), Chapter 10 of said book starts with a > quote from Sir Walter Scott: >h" > "Oh what a tangled web we weave,$ > When first we practise to deceive" >tI > The following paragraph accuses Sir Walter of "trying to deceive you by G > suggesting the word "practice" is spelled with an "s," but then againe > that was the late 1700s."b >t7 > Oops. In British English one can practise a practice.n >oI > I am going to try and review this book more objectively in the next daymH > or two. These are my first impressions only, but it ain't looking good	 > to far.a >/ > ___y > Paul Sture
 > Switzerland   & --------------0868DF7BC2855FD31DD494F1- Content-Type: text/x-vcard; charset=us-ascii;t  name="maustin.vcf"s Content-Transfer-Encoding: 7bito, Content-Description: Card for Michael Austin  Content-Disposition: attachment;  filename="maustin.vcf"    begin:vcard  n:Austin;Michael g tel;work:704-947-1089e x-mozilla-html:FALSE org:Michael Austin, Incn
 adr:;;;;;; version:2.1n+ email;internet:michaelaustininc@hotmail.comr title:Presidents x-mozilla-cpt:;0 fn:Michael E. Austin	 end:vcards  ( --------------0868DF7BC2855FD31DD494F1--   ------------------------------    Date: 28 Apr 2000 10:54:26 +0200G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>k Subject: Re: OpenVMS marketingH Message-ID: <y4hfcmfvwd.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  / Graham Allan <allan@mnhep1.hep.umn.edu> writes:-  J > True, but it seems rather... cheap... not to include it. I have multipleM > boxes, and buy media; that's fine. Some people buy single systems, in whichi3 > case $300-odd for the OS media is an aggravation.e  G Is that number correct? I remember reading something along the lines ofcB $10-20, which is OK given the cost of handling. $300 is ludicrous.   	Jan   ------------------------------   Date: 28 Apr 2000 13:26:17 GMT2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)4 Subject: Re: Porting To VMS: select function problem5 Message-ID: <8ec3hp$ap$1@mailint03.im.hou.compaq.com>f  d In article <8ean7r$19qc$1@news2atm.raleigh.ibm.com>, "Paul Nankervis" <paulnank@au1.ibm.com> writes:I :I have written a whole bunch of 'portable' TCP/IP code which also relies F :on select() - and I have been very happy with the way the VMS versionC :works.  But I recently had an example of where I needed to wake upyD :the select() mainline through a VMS AST. It would be really good toD :have some mechanism for being able to do this!!!  Perhaps even justG :some way of finding out what event flag it is using so I could set it?S  @   The select call can obviously be accessed via the TCP/IP $qio B   interface -- I've ended up using ASTs and the BG/$qio interface D   quite regularly, as it avoids a whole host of programming problemsE   for me.  (I do regularly try to avoid the use of event flag waits, eD   as I have gotten myself in trouble with this construct more times G   than it's been helpful.  I'm a regular user of EFN$C_ENF, of course.)a  E   I don't know that there are plans to alter or open up the internal cB   behaviour of the existing select call, beyond adding additional F   support for other types of file descriptors.  (This access may well D   inherently grant the sort of behaviour that you seek, but I don't =   know that will be available/documented with any certainty.)o  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Fri, 28 Apr 2000 12:36:03 -0400s5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com>h4 Subject: Re: Porting To VMS: select function problem+ Message-ID: <8ecei7$tf2$1@lead.zk3.dec.com>l  K There are two problems with select.  One is that it is implemented only fors> sockets, the other is that DECwindows isn't using sockets, andI XGetConnectionNumber does not return anything useful as input to a select 
 statement.  L We are doing some work now that "might" provide a way to hook into select(),L but this isn't a sure thing yet.  If that happens, I will be glad to look atL how to modify our X11 implementation to return something compatable that canJ be used in the select call.  But no promises right now...  this is just to. note that we are painfully aware of ths issue.  I The problem is really that some yutz embedded a truly ugly UNIX hack intogE the X11 specification and allowed the "connection number" on a "POSIXRJ compliant implementation" to return a value that could be permuted into an input to select().  L One ingenious way to solve the problem (but not reduce your porting work) isK to use the XSelectAsyncInput() or XSelectAsyncEvent() calls when you create L the window, this allows you to get AST events when an event occurs.  You canJ then use some other passing mechanism (event flag, whatever) to signal theL thread of the event.  I do this all the time (in fact, most of the time I doL a SYS$HIBER in the mainline, and do a SYS$WAKE from the AST routine I set up for all events.n  F Of course, these calls are not on all UNIX implementations ;-( -- they *were* on ULTRIX.     J Hoff Hoffman wrote in message <8ec3hp$ap$1@mailint03.im.hou.compaq.com>... >jF >In article <8ean7r$19qc$1@news2atm.raleigh.ibm.com>, "Paul Nankervis" <paulnank@au1.ibm.com> writes:J >:I have written a whole bunch of 'portable' TCP/IP code which also reliesG >:on select() - and I have been very happy with the way the VMS versionoD >:works.  But I recently had an example of where I needed to wake upE >:the select() mainline through a VMS AST. It would be really good toeE >:have some mechanism for being able to do this!!!  Perhaps even justdH >:some way of finding out what event flag it is using so I could set it? >D@ >  The select call can obviously be accessed via the TCP/IP $qioB >  interface -- I've ended up using ASTs and the BG/$qio interfaceE >  quite regularly, as it avoids a whole host of programming problemsIE >  for me.  (I do regularly try to avoid the use of event flag waits,tD >  as I have gotten myself in trouble with this construct more timesH >  than it's been helpful.  I'm a regular user of EFN$C_ENF, of course.) >rE >  I don't know that there are plans to alter or open up the internal B >  behaviour of the existing select call, beyond adding additionalF >  support for other types of file descriptors.  (This access may wellD >  inherently grant the sort of behaviour that you seek, but I don't> >  know that will be available/documented with any certainty.) > + > --------------------------- pure personaln# opinion ---------------------------o0 >   Hoff (Stephen) Hoffman   OpenVMS Engineering hoffman#xdelta.zko.dec.com >v   ------------------------------  % Date: Fri, 28 Apr 2000 08:58:44 +0100 4 From: "Thomas nilsen" <Thomas.Nilsen@kverneland.com> Subject: Print queue -> file4 Message-ID: <XAbO4.13218$By1.239098@news1.online.no>  K Is there any way to set up a print queue on OpenVMS AXP v7.1 so that printssL sent to this queue is saved as file on the system instead of going to a real printer?   Thanks in advance,  
 Thomas NilsenD
 Kverneland IT@   ------------------------------   Date: 28 Apr 2000 09:15:51 GMT3 From: gartmann@immunbio.mpg.de (Christoph Gartmann)w  Subject: Re: Print queue -> file0 Message-ID: <8ebks7$ldv$2@n.ruf.uni-freiburg.de>  k In article <XAbO4.13218$By1.239098@news1.online.no>, "Thomas nilsen" <Thomas.Nilsen@kverneland.com> writes: L >Is there any way to set up a print queue on OpenVMS AXP v7.1 so that printsM >sent to this queue is saved as file on the system instead of going to a real-	 >printer?-  , You could write a queue symbiont to do this.   Regards,    Christoph Gartmannb  H -----------------------------------------------------------------------+H | Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452 |H | Immunbiologie                                                        |H | Postfach 1169                 Internet: gartmann@immunbio.mpg.de     |H | D-79011  Freiburg, FRG                                               |H +------------ http://www.immunbio.mpg.de/english/menue.html -----------+   ------------------------------  % Date: Fri, 28 Apr 2000 12:59:46 +0200u= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>s  Subject: Re: Print queue -> file( Message-ID: <39096F22.53B0B66@gtech.com>   Thomas nilsen wrote:M > Is there any way to set up a print queue on OpenVMS AXP v7.1 so that prints-N > sent to this queue is saved as file on the system instead of going to a real
 > printer?  > Fast and dirty work-around would be to just stop the queue and look at the file.$  9 For a more sophisticated solution, then you could write ah0 print symbiont to save the file (and not print).  0 For the last possibility there are some code in:$   ftp://ftp.hhs.dk/pub/vms/symbiont//   http://www.hhs.dk/anonymous/pub/vms/symbiont/   % (it is old, but it should still work).   Arne   ------------------------------  # Date: Fri, 28 Apr 2000 09:05:51 GMTm4 From: john@laird-towers.freeserve.co.uk (John Laird)  Subject: Re: queue status values3 Message-ID: <390952e0.312104212@news.freeserve.net>.  1 On Thu, 27 Apr 2000 13:31:33 +0100, Tim Llewellyn   <tim.llewellyn@bbc.co.uk> wrote:   >3 >t >"Jean-Franois Marchal" wrote:  >  >> Bonjour ! >>I >> I'm going to use the queue status of a print queue in a DCL procedure. K >> The information displayed by a simple show queue is sufficient (stopped,c
 >> idle, ...)t >>I >> What of theses values can be set together (as AVAILABLE and STOPPED) ?lM >> What mask could be applied to the status value to get the status name in ar >> simple manner ?K >> (using separate QUEUE_IDLE, QUEUE_STOPPED, ... need to issue many getquit3 >> and consolidate the TRUE/FALSE values obtained),a >nN >I know, its a PITA, isn't it. I'd prefer one call to return the current queue >stateF >in a string, rather than all then is it stopped? is it paused?, is it >pausing?... palaver.e >r0 >I don't THINK I'm missing any $GETQUI features.  < (You've made me scurry across 6 telnet hops to find this...)  H I've used QUEUE_STATUS, which returns a longword of all status bits.  To, interpret them, you need to examine $QUIDEF.   -- g
 John Laird   ------------------------------  % Date: Fri, 28 Apr 2000 17:00:08 +0100 - From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>s" Subject: Re: So who will buy VMS ?) Message-ID: <3909B588.71DF96A2@bbc.co.uk>t   Bob Kaplow wrote:   D > [My original post on this expired, but I had asked about the price@ > difference on the same hardware configuration for VMS vs UNIX] >oF > I got a call yesterday from Karen Leonard (Manager, OpenVMS CustomerG > Programs) in response to an email I had sent to Richard Marcello. ShenM > answered several questions for me. One of them was the price difference fortM > equivalent hardware configurations with OpenVMS vs Tru64Unix. She explainedhL > that this was a concious decision on the part of Compaq to account for the? > enhanced functionality of OpenVMS vs other operating systems.h  K OK, so they admit there is enhanced functionality in the pricing structure.iG Great, thats really going to sell VMS boxes. If Compaq expect people to H shell extra for VMS because VMS is better why oh why can't they tell the" world at large that VMS is better.   Sigh  -- 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 ofp MedAS or the BBC.c   ------------------------------  % Date: Fri, 28 Apr 2000 12:22:19 -0400a5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com>t Subject: Re: SRM documentation+ Message-ID: <8ecdof$118$1@lead.zk3.dec.com>m  H No it doesn't.  Not at least in the architected parts of it, and this is certainly one.  E Documentation on the SRM callback functions can be found in the Alphan Architecture Reference Manual.  L The SRM loads the environment variables from NVRAM at power up.  You can useJ the GET/SET callbacks to read and change the *IN MEMORY* dynamic copies ofI the variables, and they will persist across a warm start (reboot) but NOTbI across a power cycle or an INIT from the console (that is, a cold start).a  J To save them permanently, you must call SAVE_ENV to write the dynamic copyF of the variables back to NVRAM.  Note you may actually need to call itK multiple times - since it may take longer than a clock tick to write all of 9 them out (see the documentation on the reference manual).m   _Fred   ? Larry Kilgallen wrote in message <2000Apr27.171258.1@eisner>... K >In article <8ea7og$dtt@gap.cco.caltech.edu>, mathog@seqaxp.bio.caltech.edui (David Mathog) writes:A >> Can anybody point me at the documentation for the SRM console?  >>E >> I'm particularly interested in understanding the SET_ENV / GET_ENVwB >> bits associated with "dispatch".   (Because I can get info, but >> "set" seems not to stick.)  >cG >I think the behavior of the SRM console varies from release to release  >and from model to model.t   ------------------------------  % Date: Fri, 28 Apr 2000 09:42:00 +0100nB From: Andrew Harrison SUNUK Consultancy <andrew.nospam@uk.sun.com>- Subject: Re: Sun is no longer the dot in .comm* Message-ID: <39094ED5.2EC0C722@uk.sun.com>   Rob Young wrote:  B > Sun is no longer ., the Root server is no an S80.  Rather ironicG > they "upgraded" from an E10000 to an S80.  Seems funny a 24 processoreH > box is viewed as more powerful than a 64 processor box.  That can't beC > can it?  Of course it can.  Wait until a certain 32 processor boxe > begins shipping ;-)o >tE > http://www.washingtonpost.com/wp-dyn/articles/A48004-2000Apr19.htmlo  - It would have been funny if it had been true.t  = However the article is incorrect, the root server was an E450t7 (small 4 CPU workgroup server) not an E10K, so the boot  is on the other foot.o  > Wow do you really need a 24 CPU S80 to replace a 4 CPU E450 !!  ? BTW it probably wasn't a 24 CPU machine, the most likely configo= would be 6 CPU's which is what the vast majority of the S80's)A installed are configured with if IDC's revenue breakdowns for ther S80 are accurate.d  B I wonder where the Washington Post got the incorrect idea that the< root server was an E10K from. No prizes for correct answers.   -- Andrew Harrisone Enterprise IT Architectn   ------------------------------  % Date: Fri, 28 Apr 2000 07:13:45 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>/ Subject: Re: TCPIP V5.0A configuration problemse7 Message-ID: <200004280713_MC2-A2F9-B32B@compuserve.com>.  D         TCPIP 5.0A was a port of the Digital Unix code rather than aC development from UCX 4.2.  So it *was* developed by Unix engineers!c  J         I suppose that having a common code base makes it easier/cheaper = to= maintain but I can't help feeling that it's a step backwards!     ? Message text written by INTERNET:winston@SSRL.SLAC.STANFORD.EDU J >In article <F1YN4.82617$2D6.2116851@news20.bellglobal.com>, "Neil Rieck"=   <n.rieck@sympatico.ca> writes: >c* >Dirk Munk <munk@home.nl> wrote in message! news:39074F3C.F726ADCC@home.nl...v	 >[ snip ]  >>J >> Somehow it shows that UCX and TCPIP were not developed by VMS engineer= sI >> :-)). >> >You got that right!J >I overheard some water cooler talk a little while back mentioning that U= CXJ >is an acronym for ULTRIX Communication eXtension. (might not be accurate=   buth
 >makes sense)s  J I think it was Ultrix Connection for VMS; the "X" is out of the "nection"=   part.s  < But that doesn't mean it was developed by Ultrix engineers.<   ------------------------------  % Date: Fri, 28 Apr 2000 08:15:27 -0500o1 From: "Dave Gudewicz" <david.gudewicz@abbott.com> / Subject: Re: TCPIP V5.0A configuration problemst8 Message-ID: <8ec2n8$gaq$1@fizban.fizban.pprd.abbott.com>  7 Another v5.0a problem observed here.  Running VMS 7.2-1s  I We recently upgraded a couple of DS20Es from 7.1-2 to 7.2-1 and along theu) way upgraded from UCX 4.x to TCPIP v5.0a.e  ; Everything seemed to go well, no problems with the upgrade.   J We noticed that a couple of TCPIP services (ftp and smtp in our case) were9 not starting at boot time, but would start just fine witha$ @TCPIP$SMTP_STARTUP.COM for example.  ? After a few go arounds with support, their recommendation were:t  % *************************************   J I have seen this type problem before and what has always resolved it is to do the following.b   set def sys$common:[sysexe]t$ rename tcpip$configuration.dat *.old rename tcpip$service.dat *.old   set def sys$manageri
 @tcpip$confige  * when it ask if you want to convert, say NO& configure the core, client, and server
 restart tcpip   H Of cource you need to know all the answers before renaming the files, so you can answer then in config.  B This is caused sometimes by corruption during the convert process.  2  *************************************************   27-APR-2000 10:29:00.00f      @  One more important thing.  When configuring the services answerE  "NO" to configure on all nodes.  Configure on each node individually:   ------------------------------  % Date: Fri, 28 Apr 2000 10:13:17 -0400 0 From: arturo saavedra <arturo.saavedra@wcom.com>/ Subject: RE: TCPIP V5.0A configuration problemsn8 Message-ID: <000501bfb11b$e67c4380$8fb024a6@wcomnet.com>  D One of the things I've found bothersome are the UNIX-type display ofE any kind of counters as oppossed to the very easy to read UCX displayh of counters.  B There is even a flaky performance problem with RSH commands using D TCPIP 5.0a that UCX did not have.  When attempting to display systemC information from a remote node using RSH command, sometimes you getrD some of the information.. at times you get part of the information.. just frustating at time.. :(           -----Original Message-----7 From: Richard B. Gilbert [mailto:DRAGON@compuserve.com]a$ Sent: Friday, April 28, 2000 7:14 AM To: Info-VAX@Mvb.Saic.Comc/ Subject: Re: TCPIP V5.0A configuration problems     D         TCPIP 5.0A was a port of the Digital Unix code rather than aC development from UCX 4.2.  So it *was* developed by Unix engineers!a  K         I suppose that having a common code base makes it easier/cheaper top= maintain but I can't help feeling that it's a step backwards!e    ? Message text written by INTERNET:winston@SSRL.SLAC.STANFORD.EDUeI >In article <F1YN4.82617$2D6.2116851@news20.bellglobal.com>, "Neil Rieck"e <n.rieck@sympatico.ca> writes: >v* >Dirk Munk <munk@home.nl> wrote in message! news:39074F3C.F726ADCC@home.nl...o	 >[ snip ]e >>J >> Somehow it shows that UCX and TCPIP were not developed by VMS engineers >> :-)). >> >You got that right!K >I overheard some water cooler talk a little while back mentioning that UCX I >is an acronym for ULTRIX Communication eXtension. (might not be accuratea bute
 >makes sense)P  I I think it was Ultrix Connection for VMS; the "X" is out of the "nection"  part.e  < But that doesn't mean it was developed by Ultrix engineers.<   ------------------------------  % Date: Fri, 28 Apr 2000 16:06:57 +0200h$ From: Paul Sture <sture.ch@sture.ch>/ Subject: Re: TCPIP V5.0A configuration problems + Message-ID: <VA.0000001e.16b289bf@sture.ch>a  8 In article <39074F3C.F726ADCC@home.nl>, Dirk Munk wrote:  > From: Dirk Munk <munk@home.nl> > Newsgroups: comp.os.vmsn. > Subject: TCPIP V5.0A configuration problems % > Date: Wed, 26 Apr 2000 20:20:02 GMTw >  > Hi,e > ' > I'm running TCPIP V5.0A on VMS 7.2-1.e > E > I am trying to change my telnet configuration a bit, but I'm havingt > trouble with that.G > For test purposes I want to log all telnet access, and I'm trying the  > configure it this way: >  > TCPIP>disable service telnet > TCPIP>set service telnet -F > /log_options=(file=sys$sysdevice:[tcpip$telnet]tcpip$telnet.log,all) > TCPIP>enable service telnetc! > TCPIP>show service telnet /fulls > I > The output shows that all log options are set, except for the log file.rA > I tried it several times, restarted TCPIP, but no luck. And thesE > directory exists of course and has the same attributes as the othero > TCPIP log directories. > D > During my attempts I made the mistake of accidentally defining the > startup file like this:v >  > TCPIP>set service telnet4 > /file=sys$sysdevice:[tcpip$telnet]tcpip$telnet.log > I > Normally this entry says 'not defined', and I would like to reset it toiJ > that situation. But there is no 'set service telnet /nofile' option, andG > /file="" will give you "" as file entry, and /file=none will give you  > 'NONE' as file entry.t > H > Hacking the services.dat file could be an option, but this is VMS, not
 > Windows.  K Yea, but it's UCX, so I'll allow myself to "hack". I've occasionally found t it the only way ;-)t   > I > Somehow it shows that UCX and TCPIP were not developed by VMS engineerse > :-)).t > : > Maybe someone has an idea how to solve these problems ?? > J I have successfully used DCL to open and read/delete the offending record ' for a user defined service in the past.t  K For some reason, we got into the position where "set service" was giving a .J duplicate record error but at the same time "set noservice" gave a record . not found error (on our user defined service).  K I much prefer the DCL approach, since trying to edit the services.dat file e% lead to at least a couple of gotchas.X  J Firstly it's an indexed file, so we do an anal/rms/fdl, edit the file and  then do something like:a  1 $ convert/fdl=ucx$services ucx$services.seq *.dat"  K That one will burn you, because ucx$services is a logical name pointing to TJ ucx$services.dat, and you need to specify /fdl=ucx$services.fdl (this had : me thinking that conv/fdl was broken until I realised...).  G Next of course, the same logical actually contains the version number,   hard coded as ;1.a  J Grr. I hate UCX. Although we are told that the new TCP/IP code taken from H Tru64 is more reliable, we are still stuck with UCX ancestry in the CLI.  B As a footnote, my spill chucker objected to "noservice" above and  suggested "nosedive". :-)e   How appropriate in the context!o ___r
 Paul Sture Switzerland    ------------------------------  # Date: Fri, 28 Apr 2000 13:50:25 GMTu% From: glendoj <glendoj@breedtech.com>g# Subject: Time Sync'ing a VMS serveri) Message-ID: <8ec4uq$qrn$1@nnrp1.deja.com>   G Is there any way to sync the time on a VMS server to another VMS server G or maybe even to a unix box?  Any help or leads would be of great help.    Thankyou   glendoji    & Sent via Deja.com http://www.deja.com/ Before you buy.r   ------------------------------   Date: 28 Apr 2000 15:21:38 GMT2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)' Subject: Re: Time Sync'ing a VMS servere6 Message-ID: <8ecaa2$3cv$2@mailint03.im.hou.compaq.com>  Q In article <8ec4uq$qrn$1@nnrp1.deja.com>, glendoj <glendoj@breedtech.com> writes:OH :Is there any way to sync the time on a VMS server to another VMS serverH :or maybe even to a unix box?  Any help or leads would be of great help.  I   This is covered in the current OpenVMS FAQ.  I'll split the informationXL   out into its own section in the next edition of the FAQ, as this question =   comes up rather often...  The OpenVMS FAQ is available via:   "     http://www.openvms.compaq.com/  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------   Date: 28 Apr 2000 14:46:06 GMT3 From: gartmann@immunbio.mpg.de (Christoph Gartmann)u' Subject: Re: Time Sync'ing a VMS server00 Message-ID: <8ec87e$1ud$1@n.ruf.uni-freiburg.de>  Q In article <8ec4uq$qrn$1@nnrp1.deja.com>, glendoj <glendoj@breedtech.com> writes:vH >Is there any way to sync the time on a VMS server to another VMS serverH >or maybe even to a unix box?  Any help or leads would be of great help.   VMS to VMS when in a cluster:o   MC SYSMAN    CONFIG SET TIME-  L VMS to Unix: if you have a TCP/IP stack use NTP. I do this with Multinet and it works great.    Regards,    Christoph Gartmannm  H -----------------------------------------------------------------------+H | Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452 |H | Immunbiologie                                                        |H | Postfach 1169                 Internet: gartmann@immunbio.mpg.de     |H | D-79011  Freiburg, FRG                                               |H +------------ http://www.immunbio.mpg.de/english/menue.html -----------+   ------------------------------  % Date: Fri, 28 Apr 2000 12:11:49 -0400e5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> 3 Subject: Re: TPU/EVE : decwindows vs character celle+ Message-ID: <8ecd4o$4dh$1@lead.zk3.dec.com>i  @ JF Mezei wrote in message <3906449E.9E714E21@vl.videotron.ca>...  D >Thanks. Sort of confirms my suspicions. Out of curiosity, which QIO
 SENSE_MODEI >flags correspond to the ability of a terminal to generate mouse events ?o >h    F None.  It is not a driver function.  I believe there is a field in theC terminal ID escape sequence that indicates that locator reports are L available, but I do not believe the terminal driver parses this into any UCB characteristics flags.  L You would have to send the escape sequence to get the ID, and then parse the	 response.u   ------------------------------  % Date: Fri, 28 Apr 2000 16:06:54 +02002$ From: Paul Sture <sture.ch@sture.ch> Subject: Re: USERNAME + Message-ID: <VA.0000001d.16b27e96@sture.ch>o  G In article <sgevgjueqtj167@corp.supernews.com>, John E. Malmberg wrote: + > From: "John E. Malmberg" <wb8tyw@qsl.net>n > Newsgroups: comp.os.vms  > Subject: Re: USERNAMEp' > Date: Wed, 26 Apr 2000 18:44:42 -0500e > 1 > Paul Sture <sture.ch#sture.ch> wrote in message ' > news:VA.00000019.0dc214c9@sture.ch...sB > > In article <009E9259.E8FECDAC@SendSpamHere.ORG>, VAXman- Brian > > Schenkenberger wrote:EF > > > Hmmm!  I wonder who would be delighted by an increase in the VMS
 > username > > > length.  ;)I > > >i1 > > $ write sys$output f$length("Schenkenberger")t > > 14 > > $e= > > Hmm. Max process name 15 bytes? What to do abount that...i > L > While 12 is now the documented maximum, it used to be possible to create aM > 13 character username, and newer versions of VMS seem to accept logins from  > it.e > L > Assuming that you could get a 14 character username into the UAF file, how, > many places would it work or break things? > M From my dim and distant experience when filename lengths were opened up back mN in VMS 4.0, it was surprising how few things broke when booting back into VMS N 3.n. Of course specifying a long filename to VMS 3 got caught by the filename N parsing routines, DIR etc didn't so much as hiccup. I think I even managed to 4 get V3 backup to restore files with V4 length names.  ( The discipline of using counted strings! ___p
 Paul Sture Switzerlands   ------------------------------  % Date: Fri, 28 Apr 2000 17:03:26 +0100s- From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>i Subject: Re: Verify of Backups) Message-ID: <3909B64E.1768029D@bbc.co.uk>u   Malcolm Dunnett wrote:  + > In article <39072B50.5CDBFB35@bbc.co.uk>,f5 >     Tim Llewellyn <tim.llewellyn@bbc.co.uk> writes:  > >MF > > I have an application in use occasionally which will not cause anyN > > explicit backup verification errors but if a backup runs concurrently withL > > it the backup is toast. Needless to say, my backup script looks for that > > job during initialization. > >tK >    Can you elaborate? Do you mean that any files open by this applicationrN > are not backed up properly? If backup doesn't report any verification errorsK > for the file then presumably it has gotten a consistant "snapshot" of thefK > files. Of course this doesn't mean the files are usable, that all depends-J > on how the application writes to them and what it expects to see in themF > when it is restarted ( of course in this case you'd also end up with7 > corrupt files if the application or system crashed ).o  H No, the database is a set of RMS files spread accross several disks. The application O in question rearranges data in this database. It is quite possible for a backupu to completeMN without verification errors (application was fiddling with a different disk at	 the time) K but still the backup is toast and it could not be restored correctly (not aa snapshot of the database).t   --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 ofn MedAS or the BBC.t   ------------------------------    Date: 28 Apr 2000 14:42:44 +0200- From: Juergen Nickelsen <nickelsen@condat.de>r< Subject: VMS/UCX responds to UDP broadcast with ICMP message, Message-ID: <87itx2iegr.fsf@chuck.condat.de>  F It has come to my attention recently that we have in our company LAN aE VAX running VMS V5.5-2 that responds to UDP broadcasts with ICMP portc unreachable messages:r  B     14:42:35.388766 mcgnum.condat.de.netbios-dgm >                7                     172.20.255.255.netbios-dgm: udp 234t7     14:42:35.389830 migra.condat.de > mcgnum.condat.de:o; 		    icmp: 172.20.255.255 udp port netbios-dgm unreachableo  A mcgnum is a PC running Linux with Samba, which is sending NetBIOS @ datagrams (UDP) to the broadcast address. This is AFAIK standardE behaviour.  migra is the VAX, and it sends a port unreachable messageoB back. This is forbidden according to RFC 1122 chapter 3 verse 2.2:  @          An ICMP error message MUST NOT be sent as the result of          receiving:r [...]1D          *    a datagram destined to an IP broadcast or IP multicast               address, or [...]   D This behaviour does not much harm, but it is quite annoying since it7 fills the logfiles of all Linux machines running Samba.s  > The VMS version is, as mentioned, V5.5-2, UCX is version 2.0B; Upgrading is not an option.e  A Not being VMS expert I tried to dig through the UCX commands (andEE managed to correctly configure DNS on the way :-) and the "DEC TCP/IPhE Services for VMS, Installation and Configuration" manual, but withoutU success so far.t  A Is there any way to disable this behaviour? Any known workaround?6   Any hint is appreciated.   -- i' Juergen Nickelsen <nickelsen@condat.de>d0 Condat AG, Alt-Moabit 91d, 10559 Berlin, Germany+ Tel. +49 30 39094-178  Fax +49 30 39094-300    ------------------------------  % Date: Fri, 28 Apr 2000 15:27:40 +0200e= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>r@ Subject: Re: VMS/UCX responds to UDP broadcast with ICMP message) Message-ID: <390991CC.EEBDC30C@gtech.com>r   Juergen Nickelsen wrote:H > It has come to my attention recently that we have in our company LAN aG > VAX running VMS V5.5-2 that responds to UDP broadcasts with ICMP portD > unreachable messages:h   What version of UCX ?X   Arne   ------------------------------  % Date: Fri, 28 Apr 2000 15:25:51 +0100 * From: "Richard Brodie" <R.Brodie@rl.ac.uk>@ Subject: Re: VMS/UCX responds to UDP broadcast with ICMP message+ Message-ID: <8ec71f$j2i@newton.cc.rl.ac.uk>n  a "Juergen Nickelsen" <nickelsen@condat.de> wrote in message news:87itx2iegr.fsf@chuck.condat.de... H > It has come to my attention recently that we have in our company LAN aG > VAX running VMS V5.5-2 that responds to UDP broadcasts with ICMP port  > unreachable messages:w ...oF > This behaviour does not much harm, but it is quite annoying since it9 > fills the logfiles of all Linux machines running Samba.o ...t@ > The VMS version is, as mentioned, V5.5-2, UCX is version 2.0B; > Upgrading is not an option.m  H To be honest, if it's a mainly cosmetic problem, live with it. Or put itA on a separate subnet to contain the problem. Of course, it shouldb@ comply with the relevant RFC's but the chances of digging up bugB fixes on 8 year old code is pretty small. A VMS historian might beF able to give you the version that a bug fix got checked in on but what@ good would it do you? There is almost zero chance of a quick fix being available. Sorry.    ------------------------------    Date: 28 Apr 2000 16:56:47 +0200- From: Juergen Nickelsen <nickelsen@condat.de>C@ Subject: Re: VMS/UCX responds to UDP broadcast with ICMP message, Message-ID: <87d7nai89c.fsf@chuck.condat.de>  , "Richard Brodie" <R.Brodie@rl.ac.uk> writes:  E > Of course, it should comply with the relevant RFC's but the chancesSC > of digging up bug fixes on 8 year old code is pretty small. A VMStD > historian might be able to give you the version that a bug fix got. > checked in on but what good would it do you?  E I wasn't exactly hoping for a bug fix, but rather for a workaround ory@ the like. (I could imagine, for instance, running a service thatE listens on udp/138 and simply reads and ignores all packets. But thatu! probably isn't worth the effort.)   > As mentioned in a previous post, we will probably go with yourC suggestion and move the machine to another (logical) network, so itrB doesn't get the broadcasts of all the SMB servers. Thanks for your help!r   -- s' Juergen Nickelsen <nickelsen@condat.de>*0 Condat AG, Alt-Moabit 91d, 10559 Berlin, Germany+ Tel. +49 30 39094-178  Fax +49 30 39094-300t   ------------------------------  % Date: Fri, 28 Apr 2000 11:29:00 -0400e0 From: arturo saavedra <arturo.saavedra@wcom.com>@ Subject: RE: VMS/UCX responds to UDP broadcast with ICMP message8 Message-ID: <000001bfb126$7a285380$8fb024a6@wcomnet.com>  ? Recall a similar problem in which the UCX broadcast address waseC incorrectly defined.  Are all broadcast addresses defined correctlyi on the servers involved?  s   -----Original Message-----4 From: Juergen Nickelsen [mailto:nickelsen@condat.de]% Sent: Friday, April 28, 2000 10:57 AMa To: Info-VAX@Mvb.Saic.Com @ Subject: Re: VMS/UCX responds to UDP broadcast with ICMP message    , "Richard Brodie" <R.Brodie@rl.ac.uk> writes:  E > Of course, it should comply with the relevant RFC's but the chancesnC > of digging up bug fixes on 8 year old code is pretty small. A VMS-D > historian might be able to give you the version that a bug fix got. > checked in on but what good would it do you?  E I wasn't exactly hoping for a bug fix, but rather for a workaround orI@ the like. (I could imagine, for instance, running a service thatE listens on udp/138 and simply reads and ignores all packets. But that ! probably isn't worth the effort.)   > As mentioned in a previous post, we will probably go with yourC suggestion and move the machine to another (logical) network, so itwB doesn't get the broadcasts of all the SMB servers. Thanks for your help!.   -- b' Juergen Nickelsen <nickelsen@condat.de>o0 Condat AG, Alt-Moabit 91d, 10559 Berlin, Germany+ Tel. +49 30 39094-178  Fax +49 30 39094-300o   ------------------------------  % Date: Fri, 28 Apr 2000 11:37:51 -0400 0 From: arturo saavedra <arturo.saavedra@wcom.com>@ Subject: RE: VMS/UCX responds to UDP broadcast with ICMP message8 Message-ID: <000101bfb127$b6932240$8fb024a6@wcomnet.com>  N Ooops scratch that.. did not noticed the part where you mentioned that VMS/UCX$ was responding with an ICMP message.         -----Original Message-----7 From: arturo saavedra [mailto:arturo.saavedra@wcom.com]-% Sent: Friday, April 28, 2000 11:29 AM2 To: Info-VAX@Mvb.Saic.Com @ Subject: RE: VMS/UCX responds to UDP broadcast with ICMP message    ? Recall a similar problem in which the UCX broadcast address was C incorrectly defined.  Are all broadcast addresses defined correctlyk on the servers involved?     -----Original Message-----4 From: Juergen Nickelsen [mailto:nickelsen@condat.de]% Sent: Friday, April 28, 2000 10:57 AM  To: Info-VAX@Mvb.Saic.Com @ Subject: Re: VMS/UCX responds to UDP broadcast with ICMP message    , "Richard Brodie" <R.Brodie@rl.ac.uk> writes:  E > Of course, it should comply with the relevant RFC's but the chances C > of digging up bug fixes on 8 year old code is pretty small. A VMSeD > historian might be able to give you the version that a bug fix got. > checked in on but what good would it do you?  E I wasn't exactly hoping for a bug fix, but rather for a workaround or @ the like. (I could imagine, for instance, running a service thatE listens on udp/138 and simply reads and ignores all packets. But that ! probably isn't worth the effort.)a  > As mentioned in a previous post, we will probably go with yourC suggestion and move the machine to another (logical) network, so ithB doesn't get the broadcasts of all the SMB servers. Thanks for your help!c   --' Juergen Nickelsen <nickelsen@condat.de>'0 Condat AG, Alt-Moabit 91d, 10559 Berlin, Germany+ Tel. +49 30 39094-178  Fax +49 30 39094-300t   ------------------------------  % Date: Fri, 28 Apr 2000 09:24:13 +0200/= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>n: Subject: Re: Where is the lexical f$getjpi("","Connectim")) Message-ID: <39093C9D.B85916A5@gtech.com>'   Terry Marosites wrote:L > I have been looking all morning for a lexical that will return a process's > current connect time,dJ > Don't really want to have to use f$CVTIME, I just can believe that there/ > isn't an f$getjpi("","Connect_time") lexical. 4 > Am I missing it or am I the only one needing this.  ? There are a F$GETJPI("","LOGINTIM") and a F$TIME() and what youo want is the difference.   ? Easy to do if it were SYS$GETJPIW and SYS$GETTIM in a program -e a bit more cumbersome in DCL.c  ) Could you rewrite it into a HLL program ?r   Arne   ------------------------------  % Date: Fri, 28 Apr 2000 10:34:50 +0100e8 From: John Macallister <J.Macallister1@physics.ox.ac.uk>: Subject: RE: Where is the lexical f$getjpi("","Connectim")L Message-ID: <F02D5A46B8AED311BE4F0090279FA2401E8332@ppnt41.physics.ox.ac.uk>  K The "SHOW PROCESS/ACC" command will give you an item "Connect time:" in its G output which you can easily access with a simple read and string searchi; sequence but you may be looking for something more compact?.   John  B Name: John B. Macallister  E-mail: j.macallister1@physics.ox.ac.ukH Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UKA Phone: +44-1865-273388 (direct)  273333 (reception)  273418 (Fax)u     -----Original Message-----6 From: Terry Marosites [mailto:TMarosites@unitedad.com]& Sent: Thursday, April 27, 2000 6:43 PM To: John Macallister6 Subject: Where is the lexical f$getjpi("","Connectim")     Hello all, -  J  Please note, I set my email to plain text, (never will turn HTML on again	 (Smile)). J I have been looking all morning for a lexical that will return a process's current connect time, H Don't really want to have to use f$CVTIME, I just can believe that there- isn't an f$getjpi("","Connect_time") lexical.a2 Am I missing it or am I the only one needing this.   Have Fun Terry    ------------------------------  % Date: Fri, 28 Apr 2000 09:27:00 -04000% From: Bob Harris <harris@zk3.dec.com>oB Subject: Re: Why does Pathworks(Mac) change mod date of .dir file?3 Message-ID: <280420000927003531%harris@zk3.dec.com>o  / [[ This message was both posted and mailed: seen;    the "To," "Cc," and "Newsgroups" headers for details. ]]n  8 In article <8eajbe$f1n$1@hecate.umd.edu>, Lawrence Bleau <bleau@umdsp.umd.edu> wrote:  L >I did a File > Get_Info on the volume after connecting to it and swiped the >comment text.  Here it is:X > M >MSAF$SERVER>  Clear Finder Invisible / by SYSTEM <26452.101.245> 27-APR-2000o >19:36C > Set Finder Invisible / by DWYER <26452.151.244> 20-APR-2000 14:08o2 > Clear Finder Invisible / by SYSTEM <26452.101.24 > P >SYSTEM is me just now connecting to the volume.  DWYER is the user who owns the >volume. > & >So, what does Finder Invisible mean? 2 >Is this something I can ask the user to change?    D If the invisible bit is set it means that the file or directory (akaF Mac folder) will not be displayed.  It will be invisible to the person sitting at the Mac.e  E Since the state of the invisible bit was changed by your just logging F in as SYSTEM _AND_ I am assuming that you did not intentionally changeC the state, I'm going to assume that the Mac you were using has some.F kind of INIT that is doing this when you connect.  (A Mac INIT is codeF that is loaded during the Mac boot process that aguments the operatingD system (like an OpenVMS loadable system service or loadable driver);D Users throw INITs on their systems to do all kinds of things (I haveG one that puts a set of eye balls on my menu bar, and about 2 or 3 dozenm other INITs that do things)).   B The first step in isolating this problem is to connect to the fileD server using different Macs.  It might be better if the Macs are notF setup by the same person so that they are less likely to have the same set of INIT's.  C Another option is to use the "Apple Menu->Control Panels->ExtensioneC Manager->Selected Set" to select just the standard Mac OS INITs andgG reboot.  See what effect this has on the invisible bit when logging in.e    E I also would try logging in as SYSTEM and as DWYER and as other users @ to see if the behavior changes depending on who logs in as well.  G See if you can find a pattern.  This might lead you to the cause of the  problem.  F If you can not isolate the problem that way, then if you want to proveF whether it is the Mac or the file server that is setting the bit, thenE you need to put an AppleTalk aware sniffer on the network and analyse G the protocol messages sent to the file server.  However, I have to tell D you that the file server does _NOT_ have any code that would set theF invisible bit, nor does the file server care about the invisible bit. G So it is unlikely that the file server would be responsible for settinguD the bit (not impossible, just highly unlikely).  This is because theD file server really doesn't care about the invisible bit.  It is justG something the Mac sets and the Finder looks at when deciding whether tom display the file or not.  B >Better yet, why is Pathworks(Mac) putting a comment on this file?  > Because customers were complaining that all of their files hadD disappeared (when viewed from the Mac), but they could see them fromE OpenVMS.  They were blaiming the file server when the file server waslF just doing what it was told to do.  Since there is no code in the fileC server to play with the invisible bit, it was decided that a way to B audit the setting of the invisible bit was needed.  Using the file@ comment was what was choosen since it placed the evidence on the affected file.  ) >Is there a way to turn this feature off?S  G If you look in the MSAF$SERVER.LOG file you should see at the beginning4, of the log an entry for the control logical 7 MSAF$SERVER_RECORD_WHO_SET_INVISIBLE.  If you create ani@ MSA$ROOT:[MSA]MSAF$SYRUN.COM procedure, you can set this controlA logical to disable the feature the next time you restart the file G server.  There should be a template file for the MSAF$SYRUN file in theu; MSA$ROOT:[MSA] directory which should help you get started.h  D Also, I think the code _ONLY_ add an invisible bit change comment ifF the file does not already have a comment _OR_ that comment begins withF the string "MSAF$SERVER>".  So try adding your own comment to the fileG that does not beging with "MSAF$SERVER>".  This should disable the filekG server from auditing that specific file.  You can use the "Get Info...".C command on the Mac or you can use the MSA$EXAMPLES:MSAF$COMMENT.EXE.2 command from OpenVMS to add a comment to the file.  A However, you should be aware, you have evidence that something iscG setting the invisible bit on your volume.  If your Mac user(s) start tocE complain that they can not see their files, you may want to re-enable,D this feature so that you can track who is setting the invisible bit.   ------------------------------  , Date: Fri, 28 Apr 2000 15:57:57 +0200 (CEST): From: "Gotfryd Smolik, VMS lists" <gotfryd@stanpol.com.pl>E Subject: [upd]Re: Does backup command backup used or allocated space?tJ Message-ID: <Pine.LNX.4.10.10004281556360.1437-100000@irys.stanpol.com.pl>  @  Minor addition to previous mail: of course, setting /BLOCK=2048E in the BACKUP command makes the mentioned save-set 8 block in lenght, ' what makes the point better to check :)p    Regards - Gotfryd   --E =====================================================================cF $ ON F$ERROR("LANGUAGE","ENGLISH","IN_MESSAGE").GT.F$ERROR("NORMAL") - 		THEN EXCUSE/OBJECT=ME . $!                        GS@stanpol.zabrze.plE =====================================================================h   ------------------------------   End of INFO-VAX 2000.237 ************************