1 INFO-VAX	Tue, 06 Dec 2005	Volume 2005 : Issue 679       Contents:& Re: Alpha uArchitecture in the news...! Re: Flushing SYS$OUTPUT log files  Re: Hobbyist VAX consoles  Re: Hobbyist VAX consoles E Re: How to convert bound volume set to single drive while  preserving D Re: How to convert bound volume set to single drive while preservingD Re: How to convert bound volume set to single drive while preservingJ Re: How to convert bound volume set to single drive while preserving linksJ Re: How to convert bound volume set to single drive while preserving linksJ Re: How to convert bound volume set to single drive while preserving linksJ Re: How to convert bound volume set to single drive while preserving linksJ Re: How to convert bound volume set to single drive while preserving linksP Re: How to convert bound volume set to single drive while preserving links links@ Re: Memory boards for the KA650 (uVAX-III) and KA655 (uVAX-III+) Miata questions. Re: Miata questions. Re: Miata questions." Re: OT: VMS and defense / missiles& Re: Pathworks network printing problem9 Re: Stupid Question: How to find speed of an Alpha's CPUs 9 RE: Stupid Question: How to find speed of an Alpha's CPUs 9 Re: Stupid Question: How to find speed of an Alpha's CPUs  Re: TLZ9L documentation? Re: V8.3 teaser 2 Re: values for vcc_ptes and show/mem/cache/full???2 Re: values for vcc_ptes and show/mem/cache/full???  F ----------------------------------------------------------------------  # Date: Tue, 06 Dec 2005 18:12:55 GMT 1 From: Keith Parris <keithparris_NOSPAM@yahoo.com> / Subject: Re: Alpha uArchitecture in the news... / Message-ID: <Hwklf.161$Iv6.90@news.cpqcorp.net>    Michael Moroney wrote:: > peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER) writes:B >>Strange. Where did EV3 live ? Not in real/sold products I think. > H > If I recall, it was a prototype only, used for Alpha code development.J > It had no floating point instructions, but other than that I don't know  > anything about it.  E Michael is correct. EV3 chips lived in ADUs (Application Development  G Units), which were in a cabinet about the size of an HSC50, and used a  ? small system running UNIX as their I/O subsystem. I joined VMS  F Engineering during the Alpha port when these were in use, just before I the first real Alpha systems (Flamingo) with EV4 chips became available.  D ADUs with EV3s had their limitations, but were much faster than the H Mannequin simulator which ran on VAX systems and could take a half-hour * just to get to the VMS banner during boot.   ------------------------------  $ Date: Tue, 6 Dec 2005 14:35:37 +08003 From: "Richard Maher" <maher_rj@hotspamnotmail.com> * Subject: Re: Flushing SYS$OUTPUT log files1 Message-ID: <dn3bhf$s42$1@news-02.connect.com.au>   	 Hi Steve,   I The other day Hoff <hoff@hp.nospam> wrote pretty much: - "The requirement  simply doesn't exist".  E (And it looks like everyone else here agrees with him. "Not so!" says  Google, but there you go.)   Having said that. . .   G >   Machine crashes?  I'd be looking very carefully at the hardware and A >   the software.  OpenVMS system crashes are just not normal nor + >   expected behaviour.  But you knew that.   L I've got VMS boxes crashing all over the place. But then that's the scenarioI I'm testing for at the moment :-) The power supply in Perth also comes in  handy as well :-(   F >   I tend to write to an application log file, and not to the defaultI >   output.  This means I have the log file when I need it, not just when F >   it gets saved to disk by a batch job or a caller.  (I also tend toE >   use commands such as the debugger SET OUTPUT, SHOW STACK and SHOW F >   CALLS, and other such activities, in order to get a better picture2 >   of the failure.  The process dump signal, too.  F Sorry, I wasn't being clear. I'm not testing why the box crashed (A: IG kicked it in the guts) I'm just proving that my software does the right K thing when a crash does happen. (The "Hoff says it'll never happen" defense # might not look good in the docs :-)   E >   (In my view, the default output device is for communications with E >   the application user, and is ill-suited for non-trivial debugging B >   and traceback information.  Now I have most certainly used theE >   default output for application tracing, and the default output is 8 >   the output of last resort for application failures.)  G Yeah, but look at all those VMS system managers and application support F people out there, bored out of their heads, staring at a screen that's7 constantly doing a $type/tail/interv=now bigjob.log :-)   H Once again, the application hasn't failed, but there is some potentiallyI usefully stuff that may not been hardened to disk. For example, let's say C the app has rejected a network connection request from someone, and J subsequently sent a curt message the the network (or application) operatorD about the fact. A more detailed message as to why the connection wasF rejected is written to the RMS file but it could be lost if the bufferI doesn't make it to disk. But more irritatingly, he has to wait interval X L before being able to see it in the log file. (Ooops! I think I'm making your6 case for you. What sort of an SLA am I dreaming of :-)  I A more useful example could be Rdb (and the RDMMON.LOG file?) anyway some K release note in the back of my mind tells me that the behaviour of a Rdb or L SQL/Services log file was changed a few years back to also flush the output.@ I could be wrong, but if Norm's around he might say yeah or nay.  K Anyway, if you're going to bring up Auditing as an option then please start J from the "What's wrong with it" point of view first. (Unless of course youJ feel that Auditing is unique in it's lack of trade-offs with regard to CPU) and I/O expense and mailbox-full issues).   G >   Current versions of OpenVMS flush the logs during $delprc and other E >   "hard" deletions; OpenVMS now invokes the rundown for rather more B >   cases.  (I don't recall which OpenVMS version implemented this >   change, off-hand.)  1 Certainly seems to work on VAX 6.2 and Alpha 7.3.   I >   Long-standing solutions for generating application log files involves I >   the use of shared output and a private application file, as described  >   in the FAQ.   H Yeah, but can you type/tail and share this "private application file" of yours?  J If you set up a private application file with all the user-friendlyness ofJ the default sys$output then surely the same flushing issues remain? What'sC the relative cost between hardening the I/O to disk in either file?   F >   During machine crashes, the only solution I am aware of is the oneE >   used by transactional systems -- you need to ensure that all data E >   is logged, or none of the data associated with the transaction is G >   logged.  RMS journaling is one solution, and most any database will ! >   provide this support as well.   J Ah so that's how transactions work? You have to put a hole in both ends of the egg. . .doh!  & >   (This is specific to transactionalF >   systems, and not to the traditional operations of applications and >   process log files.)   5 This is a single datasource run-of-the-mill log file.   . >   To get the data out, of course, $flush it.B >   (But as one programmer recently discovered, calling $flush (or@ >   fflush and fsync) after each line of output is comparatively >   expensive.)   C Who was that? fflush and fsync? Can't be anyone in this news group.   L Like everything else I suppose "it depends". If you're creating a print fileL for a phone-book then flushing every line might not be too bright. If you'reC logging network connection status then it might just be the ticket. 4 Judicious use of the "flags" argument may also help.  , >   SET OUTPUT is an approach for batch logs  L Is SET OUTPUT implemented as a $setimr AST with a $flush? If so does it keepI having to swap in a process that had been previously swapped out? It also J doesn't look too bright when it comes to working out that nothing has beenG written to the file in the last 10 years but it still keeps flushing at  interval X.   H >   SET FILE/END_OF_FILE when working with potentially-partially-written   Yep, that can help.   8 I've come round to your way of thinking. It ain't broke.   Cheers Richard  L PS. Did you go to the Technical update days in Oz? Was it cancelled? Nothing Happen?   0 "Hoff Hoffman" <hoff@hp.nospam> wrote in message( news:wQ_kf.45$Xr5.12@news.cpqcorp.net...C > In article <dn0tf9$du2$1@news-02.connect.com.au>, "Richard Maher" % <maher_rj@hotspamnotmail.com> writes:  >  > :...I have a need toK > :make sure my batch or detached jobs have successfully logged information  toE > :the disk-file (rather than just the RMS buffer) in case of machine  crash... > G >   Machine crashes?  I'd be looking very carefully at the hardware and A >   the software.  OpenVMS system crashes are just not normal nor + >   expected behaviour.  But you knew that.  >  >   As for the problem...  > F >   I tend to write to an application log file, and not to the defaultI >   output.  This means I have the log file when I need it, not just when F >   it gets saved to disk by a batch job or a caller.  (I also tend toE >   use commands such as the debugger SET OUTPUT, SHOW STACK and SHOW F >   CALLS, and other such activities, in order to get a better picture2 >   of the failure.  The process dump signal, too. > E >   (In my view, the default output device is for communications with E >   the application user, and is ill-suited for non-trivial debugging B >   and traceback information.  Now I have most certainly used theE >   default output for application tracing, and the default output is 8 >   the output of last resort for application failures.) > G >   Current versions of OpenVMS flush the logs during $delprc and other E >   "hard" deletions; OpenVMS now invokes the rundown for rather more B >   cases.  (I don't recall which OpenVMS version implemented this >   change, off-hand.) > I >   Long-standing solutions for generating application log files involves I >   the use of shared output and a private application file, as described  >   in the FAQ.  > F >   If working within device driver context (eg: if my code is what isG >   causing the system crash), I tend to use the tr_print calls.  These G >   calls manage a rinb buffer, and SDA writes this out during a system I >   crash.  See vms_macros.h and the associated SDA extension for details  >   of this. > F >   During machine crashes, the only solution I am aware of is the oneE >   used by transactional systems -- you need to ensure that all data E >   is logged, or none of the data associated with the transaction is G >   logged.  RMS journaling is one solution, and most any database will E >   provide this support as well.  (This is specific to transactional F >   systems, and not to the traditional operations of applications andC >   process log files.)  To get the data out, of course, $flush it. B >   (But as one programmer recently discovered, calling $flush (or@ >   fflush and fsync) after each line of output is comparatively >   expensive.)  > D >   SET OUTPUT is an approach for batch logs, and I also tend to useH >   SET FILE/END_OF_FILE when working with potentially-partially-writtenD >   logs.  This latter command moves the EOF, potentially exposing aA >   few more lines of output in the application log file to view.  > ( >  ---------------------------- #include' <rtfaq.h> ----------------------------- 4 >     For additional, please see the OpenVMS FAQ --  www.hp.com/go/openvms/faq , >  --------------------------- pure personal# opinion --------------------------- I >        Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[\0100]hp.com  >    ------------------------------   Date: 6 Dec 2005 07:59:37 +0100 6 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER)" Subject: Re: Hobbyist VAX consoles, Message-ID: <439544e9$1@news.langstoeger.at>  \ In article <4394BD01.E9A59043@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes:! >Peter 'EPLAN' LANGSTOEGER wrote: M >> But these babies are still useful and therefore their price hasn't dropped  >> through the floor.  > G >This is quite surprising. BTW, doing a google search for decserver did D >yield an "ad" for digital networks, which at one point, had "buy itI >direct from digital networks". But when I went on their site, I couldn't M >find prices nor ordering pages :-) I wonder how the 90M sells for brand new.   L The "new" is the DECserver 90M+ but I haven't checked the difference so far. Maybe SSH...  I >Found a really cheap 90L though. But this doesn't do TCPIP, and I am not & >sure if it comes with a power supply. > B >I have plenty of spare serial ports on my microvax II, so I don'tH >absolutely need a terminal server. However, if affordable, I certaintlyG >wouldn't mind it. But would want one that does PPP so that I could use & >it to connect my PDAs to my intranet.  ' How would a LAT only device do PPP/IP ? N Go for a 90TL and/or the more expensive ones (90M, ...) or go for a LANTRONIX.6 They are cheaper and more capable than DECservers now.   --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  % Date: Tue, 06 Dec 2005 09:35:50 -0500 , From: Javier Henderson <javier@kjsl.invalid>" Subject: Re: Hobbyist VAX consoles/ Message-ID: <dn47kc$1g55$1@stationair.kjsl.com>   C > I have plenty of spare serial ports on my microvax II, so I don't I > absolutely need a terminal server. However, if affordable, I certaintly H > wouldn't mind it. But would want one that does PPP so that I could use' > it to connect my PDAs to my intranet.   ) The 2511 I mentioned before will do that.   G The 2509 is another option, with 8 asynch ports instead of 16. In fact, = there's one on eBay right now for $30, ending in about a day.    -jav   ------------------------------  % Date: Tue, 06 Dec 2005 06:18:48 -0500 + From: Ken Robinson <kenrbnsn1@patmedia.net> N Subject: Re: How to convert bound volume set to single drive while  preserving= Message-ID: <7.0.0.16.2.20051206061028.034f4d58@patmedia.net>   - At 05:16 AM 12/6/2005, Larry Kilgallen wrote:  >In article B ><7dd80f60512051922j4a56978j66c6db3b4af587a7@mail.gmail.com>, Ken & >Robinson <kenrbnsn@gmail.com> writes:2 > > On 5 Dec 2005 20:59:27 -0600, Larry Kilgallen   > <Kilgallen@spamcop.net> wrote: > E >The initial problem stated was not changing the disk size, but going ' >from a volume set to not-a-volume-set.   C Yes, but I still believe that using "backup/noinit" will solve the  : problem of going from a bound volume set to a single disk.  E I did this once a number of years ago, so my memory might be playing  
 tricks on me.    Ken    ------------------------------   Date: 6 Dec 2005 04:16:34 -0600 - From: Kilgallen@SpamCop.net (Larry Kilgallen) M Subject: Re: How to convert bound volume set to single drive while preserving 3 Message-ID: <gY843xAkgSgu@eisner.encompasserve.org>   q In article <7dd80f60512051922j4a56978j66c6db3b4af587a7@mail.gmail.com>, Ken Robinson <kenrbnsn@gmail.com> writes: N > On 5 Dec 2005 20:59:27 -0600, Larry Kilgallen <Kilgallen@spamcop.net> wrote= > : M >> In article <4394FB16.EE1688AD@comcast.net>, David J Dachtera <djesys.nosp=  > am@comcast.net> writes:  >>9 >> I agree that it would be nice if Backup could do this.  >>H >> > I would expect BACKUP/IMAGE to take care of you. If it doesn't, andF >> > you're in a supported/supportable position, do make a bug report. >>H >> I do not believe image backup between diversely sized volume sets hasJ >> ever been supported.  Adding that capability would be much more complexG >> than the timing notifications recently discussed, and would be prone K >> to cause defects due to what has to be changed.  That which gets changed  >> gets broken.  >> > H > Using "backup/noint" is the way to restore an image backup to a largerF > disk. And now with support for expandable volumes on EVAs, you don'tD > even need to go the backup and restore route to move you data to a > bigger disk.  D The initial problem stated was not changing the disk size, but going& from a volume set to not-a-volume-set.   ------------------------------   Date: 6 Dec 2005 12:16:02 -0600 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) M Subject: Re: How to convert bound volume set to single drive while preserving 3 Message-ID: <1WGbhyoGan3e@eisner.encompasserve.org>   Z In article <11pbf56olt9mjc4@corp.supernews.com>, Dave Froble <davef@tsoft-inc.com> writes: > ? > One poster raised the issue of the file ID.  I've never used  J > multi-volume sets so do not know.  Can there be files in a multi-volume K > set with the same file ID?  If so, that could be an issue.  If not, then  * > a BACKUP/IMAGE should do what's desired. >   A    No, since the volume number is part of the FID.  The other two H    parts have a remote chance of matching.  I'm not sure /image restores5    the original FID other than certain reserved FIDs.    ------------------------------  # Date: Tue, 06 Dec 2005 12:19:37 GMT 5 From: rdeininger@mindspringdot.com (Robert Deininger) S Subject: Re: How to convert bound volume set to single drive while preserving links L Message-ID: <rdeininger-0612050719330001@user-uinj43e.dialup.mindspring.com>  E In article <1133827537.218810.48020@f14g2000cwb.googlegroups.com>, "R  Boyd" <bob@hax.com> wrote:  G >After reading the documentation and looking at the articles in Ask The : >Wizard I haven't found what I thought would surely exist. > E >If I want to convert a bound volume set and preserve links for files H >with multiple directory entries, I want BACKUP /IMAGE to do the job for >me. > C >Is there another tool or some unpublished qualifier that will tell C >BACKUP to merge all of the files from a volume set into one output ; >device?  I already know about BACKUP INPUT_SET:[*...]*.*;*  >OUTPUT_DEVICE:[*...] D >The problem with that approach is that any existing multiply linkedH >files will wind up with duplicate copies.  I'd like to have a sure-fire >way of avoiding this. > G >If BACKUP won't do this, isn't it a good idea to add the capability to H >BACKUP, since the VMS Wizard is discouraging people from creating boundF >volume sets now that significantly larger disks are available? I haveF >several volume sets I want to convert and I don't want to worry about" >messing up multiply linked files.  G I haven't used Bound Volume Sets in eons, so I'm likely forgetting some H details.  Doesn't the /NOALIAS qualifier do what you want with regard toI multiply linked files?  /NOALIAS will give you all the links but only the  original copy of the file.   ------------------------------   Date: 6 Dec 2005 06:25:21 -0800 $ From: "AEF" <spamsink2001@yahoo.com>S Subject: Re: How to convert bound volume set to single drive while preserving links C Message-ID: <1133879121.626415.150150@g49g2000cwa.googlegroups.com>    Robert Deininger wrote: G > In article <1133827537.218810.48020@f14g2000cwb.googlegroups.com>, "R  > Boyd" <bob@hax.com> wrote: > I > >After reading the documentation and looking at the articles in Ask The < > >Wizard I haven't found what I thought would surely exist. > > G > >If I want to convert a bound volume set and preserve links for files J > >with multiple directory entries, I want BACKUP /IMAGE to do the job for > >me. > > E > >Is there another tool or some unpublished qualifier that will tell E > >BACKUP to merge all of the files from a volume set into one output = > >device?  I already know about BACKUP INPUT_SET:[*...]*.*;*  > >OUTPUT_DEVICE:[*...] F > >The problem with that approach is that any existing multiply linkedJ > >files will wind up with duplicate copies.  I'd like to have a sure-fire > >way of avoiding this. > > I > >If BACKUP won't do this, isn't it a good idea to add the capability to J > >BACKUP, since the VMS Wizard is discouraging people from creating boundH > >volume sets now that significantly larger disks are available? I haveH > >several volume sets I want to convert and I don't want to worry about$ > >messing up multiply linked files. > I > I haven't used Bound Volume Sets in eons, so I'm likely forgetting some J > details.  Doesn't the /NOALIAS qualifier do what you want with regard toK > multiply linked files?  /NOALIAS will give you all the links but only the  > original copy of the file.  G While this would avoid multiple copies of such files, he would lose the B alias directory entries. If he's worried about extra copies and isE willing to manually re-create the aliases, then yes, that would work. F But he said he wanted to preserve the links. You can only do that with
 BACKUP/IMAGE.    ------------------------------  $ Date: Tue, 6 Dec 2005 16:45:17 -0000* From: "Richard Brodie" <R.Brodie@rl.ac.uk>S Subject: Re: How to convert bound volume set to single drive while preserving links 2 Message-ID: <dn4f6t$hce$1@blackmamba.itd.rl.ac.uk>  5 "Dave Froble" <davef@tsoft-inc.com> wrote in message  * news:11pbf56olt9mjc4@corp.supernews.com...  L > One poster raised the issue of the file ID.  I've never used multi-volume J > sets so do not know.  Can there be files in a multi-volume set with the M > same file ID?  If so, that could be an issue.  If not, then a BACKUP/IMAGE   > should do what's desired.   C No, because the file id is a (index, generation number, volume set) > tuple. If BACKUP/IMAGE were able to consolidate volume sets itD would have to reassign file IDs to anything not on the first volume.   ------------------------------   Date: 6 Dec 2005 09:17:00 -0800 $ From: "AEF" <spamsink2001@yahoo.com>S Subject: Re: How to convert bound volume set to single drive while preserving links B Message-ID: <1133889420.862685.74980@f14g2000cwb.googlegroups.com>   Dave Froble wrote: > AEF wrote: > > Robert Deininger wrote:  > > I > >>In article <1133827537.218810.48020@f14g2000cwb.googlegroups.com>, "R  > >>Boyd" <bob@hax.com> wrote: > >> > >>K > >>>After reading the documentation and looking at the articles in Ask The > > >>>Wizard I haven't found what I thought would surely exist. > >>> I > >>>If I want to convert a bound volume set and preserve links for files L > >>>with multiple directory entries, I want BACKUP /IMAGE to do the job for > >>>me. > >>> G > >>>Is there another tool or some unpublished qualifier that will tell G > >>>BACKUP to merge all of the files from a volume set into one output ? > >>>device?  I already know about BACKUP INPUT_SET:[*...]*.*;*  > >>>OUTPUT_DEVICE:[*...] H > >>>The problem with that approach is that any existing multiply linkedL > >>>files will wind up with duplicate copies.  I'd like to have a sure-fire > >>>way of avoiding this. > >>> K > >>>If BACKUP won't do this, isn't it a good idea to add the capability to L > >>>BACKUP, since the VMS Wizard is discouraging people from creating boundJ > >>>volume sets now that significantly larger disks are available? I haveJ > >>>several volume sets I want to convert and I don't want to worry about& > >>>messing up multiply linked files. > >>K > >>I haven't used Bound Volume Sets in eons, so I'm likely forgetting some L > >>details.  Doesn't the /NOALIAS qualifier do what you want with regard toM > >>multiply linked files?  /NOALIAS will give you all the links but only the  > >>original copy of the file. > >  > > K > > While this would avoid multiple copies of such files, he would lose the F > > alias directory entries. If he's worried about extra copies and isI > > willing to manually re-create the aliases, then yes, that would work. J > > But he said he wanted to preserve the links. You can only do that with > > BACKUP/IMAGE.  > >  > > > One poster raised the issue of the file ID.  I've never usedI > multi-volume sets so do not know.  Can there be files in a multi-volume J > set with the same file ID?  If so, that could be an issue.  If not, then* > a BACKUP/IMAGE should do what's desired.  @ Easy enough to check if you have some scratch disks to play with$ (example from a VAX/VMS 6.2 system):   $ SH DEV DISK$VOLSET  E Device                  Device           Error    Volume         Free 	 Trans Mnt F  Name                   Status           Count     Label        Blocks	 Count Cnt F $1$DKA200:     (XXXXX)  Mounted alloc        0  DISK1          2050584    1   1F $1$DKA400:     (XXXXX)  Mounted alloc        0  DISK2          2050590    1   1   $ DIREC/FILE   Directory DISK$VOLSET:[000000]   000000.DIR;1         (4,4,1) BACKUP.SYS;1         (8,8,1) BADBLK.SYS;1         (3,3,1) BADLOG.SYS;1         (9,9,1) BITMAP.SYS;1         (2,2,1) CONTIN.SYS;1         (7,7,1) CORIMG.SYS;1         (5,5,1) FILE1.;2             (13,1,1)  FILE1.;1             (11,1,2)  FILE2.;1             (12,1,2)  FILE3.;1             (13,1,2)  FILE4.;1             (11,1,1)  INDEXF.SYS;1         (1,1,1) SECURITY.SYS;1       (10,10,1) VOLSET.SYS;1         (6,6,1) WOAH.;2              (14,2,2)    Total of 16 files.  G The first number in the file id is the header number. The second is the D sequence number which having the value n indicates the nth time thisD header has been assigned to a file. The third number is the relativeE volume number. As you can see, FILE1.;1 and FILE4.;1 both have header G number 11. Two other files share number 13. Therefore, to undo a volume G set you need to re-number the file id's, both in INDEXF.SYS and all the  .DIR files!   G > The /IMAGE qualifier in BACKUP is a bit misleading.  For an identical D > copy, the /PHYSICAL qualifier would be required.  The /IMAGE is inG > reality a file structured copy, not an image (exactly the same) copy. D > Thus, copying files from a source, and getting the same files on aJ > destination, with alias directories intact, would be the result of using > the /IMAGE qualifier.   G In order to keep qualifier names short you almost have to pick slightly F misleading names. Image means a functionally equivalent copy. It makesB sense because as an "image" it "looks" the same at the "functionalE level". Physical means logical block numbers are preserved. Both will G preserve aliases (contrary to my previous posting) but neither will let E you restore a save set of a multi-member volume set to a single disk.   C One thing I noticed that I didn't expect was that you lose a little B "transparency" for file headers on the 2nd and higher volumes of aC volume set. I can use DUMP INDEXF.SYS with a carefully chosen block G number to see that block unformatted but this is not the case for files ? on rel. vol. no. 2 such as FILE3.;1. I can still do DUMP/HEADER B FILE3./NOFORMATTED, of course. Not a big deal -- just a curiosity. Comments anyone?   >  > --6 > David Froble                       Tel: 724-529-04506 > Dave Froble Enterprises, Inc.      Fax: 724-529-0596@ > DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com > 170 Grimplin Road  > Vanderbilt, PA  15486    ------------------------------   Date: 6 Dec 2005 10:35:03 -0800 $ From: "AEF" <spamsink2001@yahoo.com>S Subject: Re: How to convert bound volume set to single drive while preserving links B Message-ID: <1133894102.944378.31290@f14g2000cwb.googlegroups.com>   Richard Brodie wrote: 6 > "Dave Froble" <davef@tsoft-inc.com> wrote in message, > news:11pbf56olt9mjc4@corp.supernews.com... > M > > One poster raised the issue of the file ID.  I've never used multi-volume K > > sets so do not know.  Can there be files in a multi-volume set with the N > > same file ID?  If so, that could be an issue.  If not, then a BACKUP/IMAGE > > should do what's desired.  > E > No, because the file id is a (index, generation number, volume set) @ > tuple. If BACKUP/IMAGE were able to consolidate volume sets itF > would have to reassign file IDs to anything not on the first volume.  E You'd also have to change the relative volume numbers of FID's on the ? first volumne from 1 to 0. And all .DIR; files would have to be " modified appropriately, of course.   ------------------------------  % Date: Tue, 06 Dec 2005 11:31:49 -0500 ' From: Dave Froble <davef@tsoft-inc.com> Y Subject: Re: How to convert bound volume set to single drive while preserving links links 0 Message-ID: <11pbf56olt9mjc4@corp.supernews.com>  
 AEF wrote: > Robert Deininger wrote:  > G >>In article <1133827537.218810.48020@f14g2000cwb.googlegroups.com>, "R  >>Boyd" <bob@hax.com> wrote: >> >>I >>>After reading the documentation and looking at the articles in Ask The < >>>Wizard I haven't found what I thought would surely exist. >>> G >>>If I want to convert a bound volume set and preserve links for files J >>>with multiple directory entries, I want BACKUP /IMAGE to do the job for >>>me. >>> E >>>Is there another tool or some unpublished qualifier that will tell E >>>BACKUP to merge all of the files from a volume set into one output = >>>device?  I already know about BACKUP INPUT_SET:[*...]*.*;*  >>>OUTPUT_DEVICE:[*...] F >>>The problem with that approach is that any existing multiply linkedJ >>>files will wind up with duplicate copies.  I'd like to have a sure-fire >>>way of avoiding this. >>> I >>>If BACKUP won't do this, isn't it a good idea to add the capability to J >>>BACKUP, since the VMS Wizard is discouraging people from creating boundH >>>volume sets now that significantly larger disks are available? I haveH >>>several volume sets I want to convert and I don't want to worry about$ >>>messing up multiply linked files. >>I >>I haven't used Bound Volume Sets in eons, so I'm likely forgetting some J >>details.  Doesn't the /NOALIAS qualifier do what you want with regard toK >>multiply linked files?  /NOALIAS will give you all the links but only the  >>original copy of the file. >  > I > While this would avoid multiple copies of such files, he would lose the D > alias directory entries. If he's worried about extra copies and isG > willing to manually re-create the aliases, then yes, that would work. H > But he said he wanted to preserve the links. You can only do that with > BACKUP/IMAGE.  >   = One poster raised the issue of the file ID.  I've never used  H multi-volume sets so do not know.  Can there be files in a multi-volume I set with the same file ID?  If so, that could be an issue.  If not, then  ( a BACKUP/IMAGE should do what's desired.  F The /IMAGE qualifier in BACKUP is a bit misleading.  For an identical C copy, the /PHYSICAL qualifier would be required.  The /IMAGE is in  F reality a file structured copy, not an image (exactly the same) copy. C Thus, copying files from a source, and getting the same files on a  I destination, with alias directories intact, would be the result of using   the /IMAGE qualifier.    --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  # Date: Tue, 06 Dec 2005 17:41:40 GMT # From: hoff@hp.nospam (Hoff Hoffman) I Subject: Re: Memory boards for the KA650 (uVAX-III) and KA655 (uVAX-III+) / Message-ID: <o3klf.158$Rt6.92@news.cpqcorp.net>   i In article <1133886135.922550.69680@z14g2000cwz.googlegroups.com>, "Bob Armstrong" <bob@jfcl.com> writes: F :  Thanks for the information.  Someone else told me that all MS650-AxH :cards are M7621 and MS650-Bx are M7622, so by definition any M7622 card0 :would work with the III+. Can you confirm that?  C   The modules that meet the revision minimums are expected to work.   F   As for the root of the question, ayup, looks like it.  The M numbersI   did tend to be unique, save for a few part-populated modules I remember J   seeing.  MS650-A* does look to be M7621, and MS650-B* looks to be M7622.F   The M7621-%% and M7622-%% module variants are semiconductor supplier   variants, as I recall.  D   Do remember Google, too, as you can obviously find a wealth of old0   MicroVAX hardware information using that tool.  3   A quick check of Google found the following URLs:   2     http://cmcnabb.cc.vt.edu/dec94mds/ka650rvc.txt2     http://cmcnabb.cc.vt.edu/dec94mds/ka655x01.txt  A   ...among various other hits for a search for the keywords M7621 D   and M7622.  Better -- and faster -- than waiting for answers here,@   obviously...  I have various hardware resource pointers in the   FAQ, as well.   > :  I'll have to check my cables - I had no idea.  I upgraded aG :MicroVAX-II by swapping the CPU and memory cards for a uV-III+ and one F :M7622 card, and now I'd like to add another memory.  I actually don't7 :know which cable I've got however it seems to work OK.  : @ :  There can't be too much difference in the cables - the header* :connectors on the cards are all the same.    F   The wrong cable can and often does lead to run-time hardware errors,H   due to signal integrity issues with the speeds of the MS650-* modules.B   I had no end of problems when I upgraded an old MicroVAX, and it   tracked back to the cables.   B   If you're upgrading, also look carefully at the revisions of theC   other modules -- that was another revision problem I encountered. D   A down-rev QDSS module caused all manner of odd hangs and crashes,@   in my case, as the KA650 and KA655 could push the Q-bus ratherB   faster than the down-rev QDSS graphics controller could support.    N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq N  --------------------------- pure personal opinion ---------------------------G        Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[\0100]hp.com    ------------------------------  % Date: Tue, 06 Dec 2005 07:14:00 +0800  From: prep@prep.synonet.com  Subject: Miata questions. - Message-ID: <877jaj6ssn.fsf@prep.synonet.com>   < I have just got a 500au and a 600au. The 600 is a GL and the 500 not as I understand it.   A What is the fix for the power switch on the not GLs? It turns off  with no warning or provacation.   B Anyone know what a MAchine Check in PAL mode with a PC of c400 is?   tnx.   --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Tue, 06 Dec 2005 05:46:39 -0800 # From: "Tom Linden" <tom@kednos.com>  Subject: Re: Miata questions. ( Message-ID: <ops1c671dbzgicya@hyrrokkin>  B On Tue, 06 Dec 2005 07:14:00 +0800, <prep@prep.synonet.com> wrote:  > > I have just got a 500au and a 600au. The 600 is a GL and the > 500 not as I understand it.  > C > What is the fix for the power switch on the not GLs? It turns off ! > with no warning or provacation.  > D > Anyone know what a MAchine Check in PAL mode with a PC of c400 is? >  > tnx. >  FYI, a recent exchange       Mr. Linden:   K > BTW, David, do you know if there is a way to get a Miata to automatically J > power back up after the restoration of Power (following outage) possiblyK > through the artifice of a soldering iron?  Not currently running any of    > these J > (tho still have a GL running Tru64, doesn't have that problem) but was   > always5 > curious of any way to remedy Digital's design flaw.   F     The crude but simple way is to yank the green wire (pin 14) to theF power supply, and ground it.  If you want the existing power switch toG continue to work, it gets more complicated.  Now that I have an XP1000, F I've largely lost interest in building the gizmo I designed (but never tested) to enhance the old PWS.   H ------------------------------------------------------------------------  5     Steven M. Schweda               (+1) 651-699-9818 4     382 South Warwick Street        sms@antinode.org     Saint Paul  MN  55105-2547  I Also you can run the 500 at 600 with following switch settings (left to    right)   500       0111 1001  600       0010 0101   . I tried higher clock rates, but wouldn't boot.   Tom    ------------------------------  # Date: Tue, 06 Dec 2005 14:42:25 GMT ) From: "brian" <brian__corbin@hotmail.com>  Subject: Re: Miata questions. . Message-ID: <lrhlf.139$gp6.0@news.cpqcorp.net>  J Re turning off with out warning. Check your internal fan, If it stops the  system will power down. L Re machine check. You will need to look at the errorlog for more details on / the crash or possible non-fatal machine checks.   + anal/err/inc=(mem,cpu,bug)/out=errorlog.txt  or& dia/inc=(mem,cpu,bug)/out=errorlog.txt  K If the crashes are frequent you could try setting your console to the com1  ! serial port >>>SET CONSOLE SERIAL K and attached a PC using terminal emulation to capture the console log to a  M file. If the system crashes it may printout the machine check frame with the  J error registers.  If the system halts without saving a crashdump try this K command to extract the machine check frame from memory before rebooting or  3 initializing the system. Post the output in a reply      >>>examine -q -p 6060 -n 200   Brian       ) <prep@prep.synonet.com> wrote in message  ' news:877jaj6ssn.fsf@prep.synonet.com... = >I have just got a 500au and a 600au. The 600 is a GL and the  > 500 not as I understand it.  > C > What is the fix for the power switch on the not GLs? It turns off ! > with no warning or provacation.  > D > Anyone know what a MAchine Check in PAL mode with a PC of c400 is? >  > tnx. >  > --  > > Paul Repacholi                               1 Crescent Rd.,9 > +61 (08) 9257-1001                           Kalamunda. A >                                             West Australia 6076 , > comp.os.vms,- The Older, Grumpier Slashdot0 > Raw, Cooked or Well-done, it's all half baked.I > EPIC, The Architecture of the future, always has been, always will be.     ------------------------------   Date: 6 Dec 2005 07:58:45 -0600 B From: clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley)+ Subject: Re: OT: VMS and defense / missiles 3 Message-ID: <ddDMnoSfCaQc@eisner.encompasserve.org>   \ In article <43925D13.BF3CD24A@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes:5 > http://news.com.com/2060-10801_3-0.html?tag=nefd.bl  > J > It shows computer controlled missiles that can be purchased at Marks and > Spencer in the UK. >   K The product seems to have been discontinued as the link in the news article ( points to a product unavailable message.   Simon.   --  ; Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP I If Google's motto is "do no wrong", then how did we get Google Groups 2 ?    ------------------------------   Date: 6 Dec 2005 08:04:01 -0600 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) / Subject: Re: Pathworks network printing problem 3 Message-ID: <kZb$k4TsY767@eisner.encompasserve.org>   e In article <1133793743.176516.225310@g49g2000cwa.googlegroups.com>, johnbooth@tinyworld.co.uk writes: 2 > Hi, looking for a little help if it's out there. > I > Running OpenVMS 7..2-1 with Pathworks v5 , currently have three network F > printers, HP Laserjets, on VMS print queues available to network pcsI > using Pathworks.  Just attached a new printer, an HP Colour Jet, set up H > a queue using its tcpip address, can ping it ok, but attempts to printF > appear on the queue then disappear.  Nothing arrives at the printer. > ! > Am I missing something obvious?   A    I'd rumble through the printer's built in menus.  Amoung other @    things make sure it's set to use IP, the correct gateway, and@    whether it defaults to PCL, PS, or text.  Also verify whether    autorecognition is enabled.  H    You may need to add a formfeed after the job.  I think there are manyH    discussions of this kind of thing in the FAQ and via Google of c.o.v.   ------------------------------   Date: 6 Dec 2005 08:20:59 -0800 ' From: "Schnootling" <chuckm@dis.wa.gov> B Subject: Re: Stupid Question: How to find speed of an Alpha's CPUsC Message-ID: <1133886059.077762.296180@g44g2000cwa.googlegroups.com>   
 Hello Dan,  B I did a 'show cpu' but did not get the detailed information you've
 described.   I got something like: # CPU 0 (4 MB cache)    type 3   .... " CPU 1 (4 MB cache)    type 3   ...  > I guess I could crack open the box and look at the CPU boards.  E Oh well, thanks for the assistance. That's what I get for not backing  up my Windows toy.   Chuck    ------------------------------  $ Date: Tue, 6 Dec 2005 12:59:48 -0500' From: "Main, Kerry" <Kerry.Main@hp.com> B Subject: RE: Stupid Question: How to find speed of an Alpha's CPUsR Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB7732F6@tayexc19.americas.cpqcorp.net>   > -----Original Message-----1 > From: Schnootling [mailto:chuckm@dis.wa.gov]=20 ! > Sent: December 6, 2005 11:21 AM  > To: Info-VAX@Mvb.Saic.Com D > Subject: Re: Stupid Question: How to find speed of an Alpha's CPUs >=20 > Hello Dan, >=20D > I did a 'show cpu' but did not get the detailed information you've > described. >=20 > I got something like: % > CPU 0 (4 MB cache)    type 3   .... $ > CPU 1 (4 MB cache)    type 3   ... >=20@ > I guess I could crack open the box and look at the CPU boards. >=20G > Oh well, thanks for the assistance. That's what I get for not backing  > up my Windows toy. >=20 > Chuck  >=20 >=20   Chuck,  9 I believe Dan suggested a ">>> show config" not show cpu.      >>> help (also works)    Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  4 OpenVMS - the secure, multi-site OS that just works.   ------------------------------  # Date: Tue, 06 Dec 2005 18:57:58 GMT & From: John Reagan <john.reagan@hp.com>B Subject: Re: Stupid Question: How to find speed of an Alpha's CPUs0 Message-ID: <Wallf.166$vx6.134@news.cpqcorp.net>   Main, Kerry wrote: >  >  > Chuck, > ; > I believe Dan suggested a ">>> show config" not show cpu.  >  >   7 Also, an SDA CLUE CONFIG should show the speed as well.      --   John Reagan / HP Pascal/{A|I}MACRO for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------   Date: 6 Dec 2005 06:01:43 -0800 1 From: "hanblo@netscape.net" <hanblo@netscape.net> ! Subject: Re: TLZ9L documentation? C Message-ID: <1133877702.964322.311110@g43g2000cwa.googlegroups.com>    Hello, could this be of interest?  ( http://www.tbcs-online.de/doku/tlz09.pdf   Regards  Hans   ------------------------------  % Date: Tue, 06 Dec 2005 03:16:43 -0500 ' From: Dave Froble <davef@tsoft-inc.com>  Subject: Re: V8.3 teaser0 Message-ID: <11pai4s9gg78ed9@corp.supernews.com>   John Santos wrote: > JF Mezei wrote:  >  >> H Vlems wrote:  >>L >>> It's perfectly readable, but I guess the real question is how often have% >>> system managers needed to compute ; >>> how long that d***d backup process is going to take....  >> >> >> >>J >> Thing is that whenever I do something serious with backup, it is alwaysK >> from a command procedure that is run at batch so I have log files and so C >> it doesn't tie up a terminal forever. And since it may be making C >> requests for tapes etc, having a free terminal to reply to those  >> requests is handy :-) >  > C > Yeah, a major exception being when you are sitting at the console A > with lots of irate users breathing down your neck as you try to @ > restore the system disk while booted from the CD-ROM, or otherG > random crisis restore.  (Since it was a crisis, it was unpredictable, B > so you don't *have* a procedure that does exactly what you want.  E Yeah!  Now you're talking about my world.  The greatest benefit from  C this idea is when you need to do something unique, and everyone is  B asking you "how long" every 30 seconds, and you don't have a clue.  K I only get involved in a restore and such after the s**t is knee high.  :-)   E > Routine backups (and retrievals, if you are using your backup tapes D > as a poor-man's archive system), which are all scripted and run asF > batch jobs in the background, wouldn't benefit from this as there isF > not keyboard to type ctrl/T on, and anyway you already know how long > they take. > G > The reason I don't script and submit as batches emergency backups and E > restores is I'm never exactly sure what I need to do until I do it, ? > and you get much faster feed-back about mistakes if you do it D > interactively.  (If it is something you need to do more than once,B > you can always cut-and-paste from the recall buffer to create anI > appropriate command procedure once you've got the problem nailed down.)  >  >>H >> The big question. If Backup is to calculate estimated time remaining,F >> Guy should be aware that on VAX (especially a Microvax II with TK50G >> backing up 10 gigs), the estimated time might be in weeks instead of  >> hours :-) >  > 3 > VMS time format is good for another 7994 years...  >  >> >>E >> What would be really neat is if one could interrogate a BACKUP job J >> running in batch and get a reply back from it giving the equivalent of 
 >> CTRL-T. >> > G > You got me thinking the exact same thing when reading the 1st part of  > your post... > K >> Perhaps something like REQUEST/TO=BACKUP STATUS   and then, BACKUP would ' >> reply with the relevant information.  >  > G > Might need a different mechanism, since backup (I think) only listens D > for OPCOM messages when it is waiting for something, like the nextD > volume to be mounted.  Also, you would probably want to be able toG > specify which backup process; there could be several of them running.  > D > But some way to trigger BACKUP to send the status info as an OPCOMA > message (can it do this from an AST?) might be the simplest and E > cleanest method, since it would use mostly existing infrastructure.  > D > (REPLY/ENABLE=BACKUP to see the messages, one of the many terminalG > broadcast interceptors or OPCOM listeners could grab them if needed.)  >      --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  $ Date: Tue, 6 Dec 2005 07:02:24 -0500 From: "Jilly" <jilly@hp.com>; Subject: Re: values for vcc_ptes and show/mem/cache/full??? , Message-ID: <43957dd1$1@usenet01.boi.hp.com>   The article at  V http://h18000.www1.hp.com/support/asktima/operating_systems/CHAMP_SRC940308008176.htmlB has the information that you need concerning tuning VIOC on a VAX.    3 "syslost" <wm.reynolds@gmail.com> wrote in message  = news:1133817292.919658.138620@g14g2000cwa.googlegroups.com... I > So, if I have 2 gig memory (which we do) on my VAX 7000, which show mem I > reports to be 4,194,304 total pages.  And vcc_ptes is set to 10,000,000 F > pages (5.12 gig), more memory than I have could be allocated to VIOCC > (causing hard faults) ?  Should vcc_ptes be set to a lower value? G > Autogen wants the value to be even higher, ("The calculated value was  > 2,000,000,000"). >    ------------------------------   Date: 6 Dec 2005 07:01:33 -0800 ' From: "syslost" <wm.reynolds@gmail.com> ; Subject: Re: values for vcc_ptes and show/mem/cache/full??? B Message-ID: <1133880898.016548.78180@z14g2000cwz.googlegroups.com>  ( Thanks for the pointer to the article!!!   ------------------------------   End of INFO-VAX 2005.679 ************************                                                                                                                                                                                                                    @-n]Pn}JڠˏOL6%r&鵚о0N!5 L\Nk\A5=li Bn=R u&ai*-1Ak= d(HsRd0\ԗu&Pj%t\}
Ȗ0&?54SgMStA1- \ F]>J?n^a0nv؆\xnP
(s5ZIax,?.F}z-(B
XhCcbIao

XtB(	с@x
eК
t0li8t@:՗CG.{sTs[@M&ہ z\g.B
2ft54V	c\Մ	٤/[6}р𚴞E! >A胂@	8B@ #}1PAr0_(0AĀa8 z4&ۥq4$4]58
AUY\-<ކKx`ph 1E}*i
M9fٳ{b@`qG8_܂g׬5?R{<wp,R6M+ݫna_Wo`@Ռ#[w2Y
svXu$XXA3XA7n#){jח
vD(WzCP`k
+BY@!C[z2J
P6APQ ԑ(Aqtf z6݆Ѐ'$r`q,k+
\PK
םU
MoENDmY]zaS AA-U ؖ
A(ʅ
űz]d(Fch"?T(Hpټ!
G	뗼n riUC,Dl\E-CAm"