1 INFO-VAX	Thu, 13 Jul 2006	Volume 2006 : Issue 386       Contents: Re: 64K limit for socket send  Re: 64K limit for socket send  Re: 64K limit for socket send  Re: Alpha remembrance day  Re: Alpha remembrance day  Re: Alpha remembrance day  Re: Alpha remembrance day  Re: File Conversion Problem  Re: File Conversion Problem  Re: Hoffman LabsG Re: How to create a new user account: Request Example Of How To Do This P Re: How to create a new user account: Request Example Of How To Do This ThisThis4 Number of consecutive failures allowed in batch job?8 Re: Number of consecutive failures allowed in batch job?8 Re: Number of consecutive failures allowed in batch job?8 Re: Number of consecutive failures allowed in batch job?8 Re: Number of consecutive failures allowed in batch job?8 Re: Number of consecutive failures allowed in batch job?8 Re: Number of consecutive failures allowed in batch job?8 Re: Number of consecutive failures allowed in batch job?8 Re: Number of consecutive failures allowed in batch job? Re: OpenVMS V8.3 is in sight Re: OpenVMS V8.3 is in sight Re: Oracle10g on IA64 8.2-1  Re: VAX 66x0 & CHARON VAX  Re: VAX 66x0 & CHARON VAX $ Re: Where Are The System Directories$ Re: Where Are The System Directories  F ----------------------------------------------------------------------    Date: 12 Jul 2006 13:00:47 -0500 From: briggs@encompasserve.org& Subject: Re: 64K limit for socket send3 Message-ID: <gSeC9FdG1pWW@eisner.encompasserve.org>   _ In article <Yg8tg.391$eB6.302@news.cpqcorp.net>, Hoff Hoffman <hoff-remove-this@hp.com> writes:  > Dave Froble wrote:" >> briggs@encompasserve.org wrote: > > >>> i.e. If you have 80K to send, send 16K first and then 64K. >>  H >> I've never heard of doing this before this thread.  Why would one do K >> this instead of just sending chunks of the message until it's all sent,  4 >> with the last chunk being whatever size required?  = I'm piggy backing on Hoff's reply since I haven't seen Dave's  posting yet.  A If you go back upthread a few messages you'll find mention of the F "Nagle" algorithm.  This is an optimization to TCP that tries to avoidF sending lots of little TCP segments, instead "saving up" and sending a few full ones instead.  D Remember the situation we have here.  The TCP stack has a limitationA on the maximum number of bytes we can put in a single "send" call " from our application to the stack.  E We're in control of the application.  We're not in control of the TCP D stack.  We have a bunch of data to send and it's more data than will fit into a single "send" call.  E So we split it up and use multiple "send"s.  Nothing wrong with that. C TCP is a byte stream protocol.  All that matters is the sequence of  bytes in the stream.  C The way Nagle works (review upthread for a much better description) B is that if our last "send" amounts to less than a full packet, theC TCP stack will save it in its buffer and wait for us to "send" some D more.  We don't want that.  We want the TCP stack to send the packet now.  H So if our TCP stack's send limit is 64,000 bytes and we have 64001 bytes0 we send 1 byte then 64000 bytes and evade Nagle.  ? If we had sent 64001 bytes then 1 byte, our 1 byte would not be > transmitted immediately, even if the partner's TCP window size
 permitted it.   H The original suggestion was to make sure the final send was greater thanI the connection MSS.  My suggestion was simpler.  Make sure the final send 5 is equal to the send limit imposed by your TCP stack.    ------------------------------  % Date: Wed, 12 Jul 2006 20:01:57 -0400 ' From: Dave Froble <davef@tsoft-inc.com> & Subject: Re: 64K limit for socket send9 Message-ID: <ANOdnWuav4CmESjZnZ2dnUVZ_t-dnZ2d@libcom.com>    briggs@encompasserve.org wrote: a > In article <Yg8tg.391$eB6.302@news.cpqcorp.net>, Hoff Hoffman <hoff-remove-this@hp.com> writes:  >> Dave Froble wrote: # >>> briggs@encompasserve.org wrote: ? >>>> i.e. If you have 80K to send, send 16K first and then 64K. I >>> I've never heard of doing this before this thread.  Why would one do  L >>> this instead of just sending chunks of the message until it's all sent, 5 >>> with the last chunk being whatever size required?  > ? > I'm piggy backing on Hoff's reply since I haven't seen Dave's  > posting yet. > C > If you go back upthread a few messages you'll find mention of the H > "Nagle" algorithm.  This is an optimization to TCP that tries to avoidH > sending lots of little TCP segments, instead "saving up" and sending a > few full ones instead. > F > Remember the situation we have here.  The TCP stack has a limitationC > on the maximum number of bytes we can put in a single "send" call $ > from our application to the stack. > G > We're in control of the application.  We're not in control of the TCP F > stack.  We have a bunch of data to send and it's more data than will  > fit into a single "send" call. > G > So we split it up and use multiple "send"s.  Nothing wrong with that. E > TCP is a byte stream protocol.  All that matters is the sequence of  > bytes in the stream. > E > The way Nagle works (review upthread for a much better description) D > is that if our last "send" amounts to less than a full packet, theE > TCP stack will save it in its buffer and wait for us to "send" some F > more.  We don't want that.  We want the TCP stack to send the packet > now. > J > So if our TCP stack's send limit is 64,000 bytes and we have 64001 bytes2 > we send 1 byte then 64000 bytes and evade Nagle. > A > If we had sent 64001 bytes then 1 byte, our 1 byte would not be @ > transmitted immediately, even if the partner's TCP window size > permitted it.  > J > The original suggestion was to make sure the final send was greater thanK > the connection MSS.  My suggestion was simpler.  Make sure the final send 7 > is equal to the send limit imposed by your TCP stack.   ( Ok, I'm aware of most of what you write.  G In the socket communications applications I've designed, the following   happens:  / 0) service is listening for connection requests  1) client requests connection  2) Service grants connection 3) Client sends request , 4) Service performs whatever, and sends data1 5) client receives data and closes the connection 9 6) service goes back to listening for connection requests   G It fits the application requirements well.  Not much call for multiple  E requests and replies from the client.  I'm aware that there could be  3 applications that differ widely from such a design.   E So my question is, say the data is 390,000 bytes, and my buffer used  H within the application is 25,000 bytes.  The applications have send and ; receive functions that will send/receive until all data is  G sent/received.  I declare data size at the front of the data stream to  H insure all is sent/received, and to know when no more is to be expected.  H The send function will keep parsing off 25K bytes and sending them in a G loop until all the data is sent.  The receive function will read until  I it has the required number of bytes.  This seems to work well.  For more   than a few years anyway.  A Now you're making me wonder whether I've introduced unneeded and  ' unwanted delays into the transmissions.   G Once I finish sending data, how long will TCP/IP take to send the last  E partial buffer.  Note that I'm assuming that my buffer and that used  - internal to TCP/IP will not be the same size.   F Should I be attempting to match size with the internal TCP/IP buffers?   Is it worth the effort?   G I sort of like it better as a 'black box' that just works, rather than  1 having to always customize the size of the sends.    --  4 David Froble                       Tel: 724-529-0450> Dave Froble Enterprises, Inc.      E-Mail: davef@tsoft-inc.com DFE Ultralights, Inc.  170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  # Date: Thu, 13 Jul 2006 01:27:25 GMT % From: Rick Jones <rick.jones2@hp.com> & Subject: Re: 64K limit for socket send/ Message-ID: <1khtg.427$m_6.62@news.cpqcorp.net>   ( Dave Froble <davef@tsoft-inc.com> wrote:  I > In the socket communications applications I've designed, the following  
 > happens:  1 > 0) service is listening for connection requests  > 1) client requests connection  > 2) Service grants connection > 3) Client sends request . > 4) Service performs whatever, and sends data3 > 5) client receives data and closes the connection ; > 6) service goes back to listening for connection requests   E If the server is the first to initiate close then that is an implicit F flush of any data queued awaiting Nagle.  That is what things are/were& like for HTTP 1.0.  So, that would be:  , 4) Service performs whatever, and sends data' 4.a) Service performs shutdown(SHUT_WR)   D 4.b) Client receives last byte of data and the shutdown notificationB      (read return of zero triggered by the receipt of the server's      FINishied segment)    5) Client calls close() H 6) Service receives the indication of client close with a zero-byte read 7) Serice calls close()  etc etc   I > It fits the application requirements well.  Not much call for multiple  G > requests and replies from the client.  I'm aware that there could be  5 > applications that differ widely from such a design.   G > So my question is, say the data is 390,000 bytes, and my buffer used  J > within the application is 25,000 bytes.  The applications have send and = > receive functions that will send/receive until all data is  I > sent/received.  I declare data size at the front of the data stream to  J > insure all is sent/received, and to know when no more is to be expected.  J > The send function will keep parsing off 25K bytes and sending them in a I > loop until all the data is sent.  The receive function will read until  K > it has the required number of bytes.  This seems to work well.  For more   > than a few years anyway.  C > Now you're making me wonder whether I've introduced unneeded and  ) > unwanted delays into the transmissions.   > In many respects it is a race.  And often as not, if the totalA transfer time is rather longer than the receiver's standalone ACK ? timer, any delay for a last, sub-MSS send (and the send must be E sub-MSS to hit this) might be epsilon. And it may not be a standalone B ACK - the data stream could be such that the last full-MSS segmentD before the last sub-MSS send was enough to induce a window update or# an immediate ACK from the reciever.   ' Broad handwaving of when ACKs are sent:   * a) piggy-backed on data goin the other way5 b) piggy-backed on window updates going the other way A c) after the standalone ACK timer expires (50 to 200 milliseconds     depending on the stack)  I > Once I finish sending data, how long will TCP/IP take to send the last  G > partial buffer.  Note that I'm assuming that my buffer and that used  / > internal to TCP/IP will not be the same size.   C _IFF_ the last send is sub-MSS it will take no longer than it takes @ for the remote reciever to send an ACK, which will depend on the above.  I > I sort of like it better as a 'black box' that just works, rather than  3 > having to always customize the size of the sends.   F I've been hesitant to mention it because it is soooo often abused as aC way to kludge around application bugs, but there is the TCP_NODELAY E setsockopt() which will disable the Nagle algorithm for a connection.   D The abuse is generally when applications send "logically associated"A data in separate calls to send() - the example I often give is an F email client sending the email headers separately from the email data.< For a more contemporary exmaple replace "email" with "http."  A Typically in such applications, both the app headers and data are D small and likely to be affected by Nagle.  The proper fix then is toE use a gathering write() like writev() or sendmsg().  Lazy programmers C will sometimes just set TCP_NODELAY.  They see this big increase in E wall-clock perf but don't realize they are sending many more segments < than they should and so consuming more CPU than they should.  
 rick jones --  ? Process shall set you free from the need for rational thought.  F these opinions are mine, all mine; HP might not want them anyway... :)D feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...   ------------------------------  % Date: Wed, 12 Jul 2006 19:26:50 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> " Subject: Re: Alpha remembrance day, Message-ID: <44B58532.7E747F91@teksavvy.com>   etmsreec@yahoo.co.uk wrote: I > The cost justifications for releasing 8.x on VAX are such that it isn't  > commercially viable.      G The cost of NOT doing the 8.x on VAX is HP's credibility and total lack E of respect for any future promises/commitments they make. You'd think C that they would have held that one commitment to try to rebuild the ; distrust as the result of premeditated euthanasia of Alpha.    ------------------------------  % Date: Wed, 12 Jul 2006 20:05:56 -0400 ' From: Dave Froble <davef@tsoft-inc.com> " Subject: Re: Alpha remembrance day9 Message-ID: <ANOdnWqav4C1ECjZnZ2dnUVZ_t-dnZ2d@libcom.com>    Michael Kraemer wrote:P > In article <1152716151.893722.152140@p79g2000cwp.googlegroups.com>, "Ed Wilts" > <ewilts@ewilts.org> writes: E >> At Decus Cincinnati 1992, I saw a Galaxy system with VMS and Linux 0 >> running at the same time.  Now that was cool! >  > huh ? galaxy in 1992 ?: > And Linux ? Was barely usable on intel, let alone alpha.  I I don't remember the year, but before Galaxy was released, the engineers  1 were showing a system running as described above.    --  4 David Froble                       Tel: 724-529-0450> Dave Froble Enterprises, Inc.      E-Mail: davef@tsoft-inc.com DFE Ultralights, Inc.  170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Wed, 12 Jul 2006 18:10:37 -0700 # From: "Tom Linden" <tom@kednos.com> " Subject: Re: Alpha remembrance day) Message-ID: <op.tclr7zlkzgicya@hyrrokkin>   H On Wed, 12 Jul 2006 17:05:56 -0700, Dave Froble <davef@tsoft-inc.com>  =   wrote:   > Michael Kraemer wrote:I >> In article <1152716151.893722.152140@p79g2000cwp.googlegroups.com>, "=  Ed  =   	 >> Wilts"  >> <ewilts@ewilts.org> writes:F >>> At Decus Cincinnati 1992, I saw a Galaxy system with VMS and Linux1 >>> running at the same time.  Now that was cool!  >>  huh ? galaxy in 1992 ?; >> And Linux ? Was barely usable on intel, let alone alpha.  > I > I don't remember the year, but before Galaxy was released, the enginee=  rs  =   3 > were showing a system running as described above.  > I I checked with the alpha linux guys (Thinking more about the port I saw =  was  at DECUS 1995 in Washington DC)     3 On 12 Jul, 2006, at 13:54, Michal Jaegermann wrote: = > On Wed, Jul 12, 2006 at 08:14:30AM -0700, Tom Linden wrote: I >> Can someone tell me when the first port of Linux to Digital HW ocurre=  d? > ? > As far as I know the first Alpha port was done in May 1994 by E > Jim Paradis from DEC labs in Massachusetts.  This was a 32-bit port @ > but pretty quickly this was replaced by a 64-bit version which > become "the one".  > - >> And I assuming that would have been Alpha.  > D > You wonder about MIPS?  I do not know what was a timeline on that. > + >> I recall seeing a port at  Decus (1994?)   D It was done AT the DECUS Symposium, in Memphis as I recall, by Linus& himself with the help of the Unix SIG.  I > That was likely a BLADE release (64-bit) which happened later in 1994.=    > ; >> In fact, I even got a replica NH license plate from John  >> 'Mad Dog' Hansen. > ? > Well, probably Jon 'maddog' Hall (who, among other things got = > DEC to send a machine to Linus) but other than that ... :-)   I Back in those days, DEC was all VMS except for a couple of "oddballs" --=   I "maddog" being one on the inside and the rest of the Unix SIG comprising=   C the rest.  Jon invited Linus to DECUS to talk about his project and  we allG had lots of fun drinking our way down some river on a paddle wheel, and 1 talking about how Unix would one day replace VMS.   G Jon then managed to convince DEC marketing folks (those with the budget F dollars) to provide Linus with one of the first Alpha machines so that% he could do his porting work on that.      T.T.F.N. William H. Magill    ------------------------------  % Date: Wed, 12 Jul 2006 22:36:35 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> " Subject: Re: Alpha remembrance day, Message-ID: <44B5B1A0.1FDDE686@teksavvy.com>   Dave Froble wrote:H >> >> At Decus Cincinnati 1992, I saw a Galaxy system with VMS and Linux  J > I don't remember the year, but before Galaxy was released, the engineers3 > were showing a system running as described above.     A In 1992, VMS was barely capable of running on Alpha (wasn't it at  version 1.6) ?      D My recollection is that Galaxy was launched  towards the very end ofF Digital or early Compaq circa 1999.  The DECUS demonstration was on anM alpha server 4100 with special firmware before the real GS machines were out.   @ As I recall, Linux was chosen because it was able to behave in aD multi-instance machine with regards to respecting which devices were- assigned to it and not touching other dvices.      GS machines came when EV6 came.    ------------------------------    Date: 12 Jul 2006 12:58:29 -0700  From: "Ian Miller" <ijm@uk2.net>$ Subject: Re: File Conversion ProblemA Message-ID: <1152734309.679443.49020@35g2000cwc.googlegroups.com>   E It may be that the file attributes were altered but not the contents, A in whch case SET FILE/ATTR can be used to correct the attributes. D CONVERT can change the file contents as part of the conversion which may not be needed.   ------------------------------    Date: 12 Jul 2006 15:37:27 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) $ Subject: Re: File Conversion Problem3 Message-ID: <2mBqlMgz3JKO@eisner.encompasserve.org>    In article <1152722413.364010.157440@h48g2000cwc.googlegroups.com>, "Tim ffrench-Lynch" <tim.newsgroups.account@googlemail.com> writes: C > A colleague has a file that is supposed to be an ordinary fortran E > source code file as produced by edit/edt with variable line length. F > However at some point in the past the file has been put through someF > sort of archiving process that has brought it back as 512 byte fixedC > length records with CR LF characters within the 512 byte records.  > G > Using Convert/FDL I can convert the file to variable length with line @ > breaks in place of each CR LF, but the line breaks at 512 byte9 > intervals are preserved leading to odd breaks in lines.  >   > Any ideas how to fix the file?  D    You don't want to CONVERT the file, that just changes the alreadyD    inconsistent meta data to new and possibly worse meta data.  You 2    want to tell RMS what the meta data already is.  *    $ set file/attr=rfm:stm <original-file>   ------------------------------  % Date: Wed, 12 Jul 2006 21:44:22 -0700 % From: Crabs <IHateSpam@SpamSucks.com>  Subject: Re: Hoffman Labs : Message-ID: <3s2dnYAV6bU7UijZnZ2dnUVZ_sadnZ2d@comcast.com>   syslost wrote:4 > Only one person can really confirm of deny this... >  >  > Ian Miller wrote:  > F >>the openvms.org newsletter of 20-MAY-2006 reported this site had new
 >>content. >  > - And he's being suspiciously quiet about it...    TomC   ------------------------------    Date: 12 Jul 2006 12:43:04 -0700  From: "Ian Miller" <ijm@uk2.net>P Subject: Re: How to create a new user account: Request Example Of How To Do ThisC Message-ID: <1152733384.655728.213140@p79g2000cwp.googlegroups.com>   & As you appear to be new at OpenVMS see. http://www.openvms.org/pages.php?page=Beginner  
 in particular , http://h71000.www7.hp.com/wbt/pc/welcome.htm   for some intro sysman material   ------------------------------  % Date: Wed, 12 Jul 2006 13:34:14 -0400 3 From: "Richard B. Gilbert" <rgilbert88@comcast.net> Y Subject: Re: How to create a new user account: Request Example Of How To Do This ThisThis : Message-ID: <qqqdnZJ_4N0FryjZnZ2dnUVZ_radnZ2d@comcast.com>   Chris L wrote:  I > Can you give me all the steps that I would have to go through to create  > a new user account.  >  > Thank you, > Christopher Lusardi  >   3 Log in as SYSTEM (or some other privileged account)  $ SET DEFAULT SYS$SYSTEM $ RUN AUTHORIZE 	 UAF> HELP  ... 
 UAF> HELP ADD  ...    That should get you started.  7 You need to read a whole bunch of manuals.  Start with: A http://h71000.www7.hp.com/doc/732final/6048/6048pro_contents.html   E There are manuals for just about everything somewhere near the above.   E What you really need is about 80 hours of classroom instruction: VMS  I System Management Part I and VMS System Management Part II.   It used to  G be one course twenty years ago but things have gotten more complicated  I or the students have gotten dumber.  They probably cost about $2500 each  I so you need an employer to pay for them, or you need to be independently   wealthy!  
 Good luck.   ------------------------------    Date: 12 Jul 2006 12:23:14 -0700 From: alyciac@aol.com = Subject: Number of consecutive failures allowed in batch job? A Message-ID: <1152732194.011468.9030@s13g2000cwa.googlegroups.com>   C I'm running a VMS 6.1 system.  Last weekend, I had a batch job stop G after trying and failing to delete 24 files.  (This is not a protection F issue - the files just were not there to be deleted)  There is nothingG in my error trapping that calls for such an action.  Is there some sort D of maximum number of consecutive errors, after which the system willF terminate a job?  The batch job in question is a series of DCL commandD procedures which execute one after another.  After the proc with theE deletes failed, the 'parent' proc continued with its cleanup and sent ? mail announcing the failure, so the process didn't simply stop.    Any thoughts would be welcome!   ------------------------------  % Date: Wed, 12 Jul 2006 15:51:47 -0400  From: norm.raphael@metso.comA Subject: Re: Number of consecutive failures allowed in batch job? Q Message-ID: <OFA6A597DA.CD91D9C7-ON852571A9.006CFD40-852571A9.006D1C9D@metso.com>   	 $ HELP ON    and    $ HELP  SET  ON        Example             $ SET NOON         $ DELETE  *.SAV;*        $ SET ON         $ COPY  *.OBJ  *.SAV   E            This command procedure routinely copies all object modules E            into new files with the file type .SAV. The DELETE command G            first deletes all existing files with the .SAV file type, if I            any. The SET NOON command ensures that the procedure continues G            executing even if there are no files with the .SAV file type J            in the current directory. Following the DELETE command, the SETJ            ON command restores error checking. Then the COPY command makes<            copies of all existing files with .OBJ file type.  0 alyciac@aol.com wrote on 07/12/2006 03:23:14 PM:  E > I'm running a VMS 6.1 system.  Last weekend, I had a batch job stop I > after trying and failing to delete 24 files.  (This is not a protection H > issue - the files just were not there to be deleted)  There is nothingI > in my error trapping that calls for such an action.  Is there some sort F > of maximum number of consecutive errors, after which the system willH > terminate a job?  The batch job in question is a series of DCL commandF > procedures which execute one after another.  After the proc with theG > deletes failed, the 'parent' proc continued with its cleanup and sent A > mail announcing the failure, so the process didn't simply stop.  >   > Any thoughts would be welcome! >    ------------------------------  % Date: Wed, 12 Jul 2006 15:51:57 -0400 3 From: "Richard B. Gilbert" <rgilbert88@comcast.net> A Subject: Re: Number of consecutive failures allowed in batch job? : Message-ID: <EMGdne7Q9pZCzyjZnZ2dnUVZ_ridnZ2d@comcast.com>   alyciac@aol.com wrote:  E > I'm running a VMS 6.1 system.  Last weekend, I had a batch job stop I > after trying and failing to delete 24 files.  (This is not a protection H > issue - the files just were not there to be deleted)  There is nothingI > in my error trapping that calls for such an action.  Is there some sort F > of maximum number of consecutive errors, after which the system will > terminate a job?    H Not that I know of.  The default is to exit on the first error.  If you G want to avoid the errors, you could test for the existance of the file   before trying to delete it.   H Or, if the nonexistance of the file is an error, you can trap the error  with ON ERROR THEN .... $ and try to do something intelligent.  F If you want to ignore the error, you can either SET NOON or code your F error handling to ignore the RMS-E-FNF; you'll have to grub around to H find the numeric code for the message, check $STATUS for that code, and F resume processing.  If you want to get really fancy, you can mask off I the low order three bits of $STATUS so as to trap F, E, and W severities.   I DCL allows you to do just about anything you want if you want to work at   it hard enough.    <snip>   ------------------------------  % Date: Wed, 12 Jul 2006 16:08:42 -0400 * From: "Syltrem" <syltremzulu@videotron.ca>A Subject: Re: Number of consecutive failures allowed in batch job? 0 Message-ID: <12balipc3funu11@corp.supernews.com>  ? "Richard B. Gilbert" <rgilbert88@comcast.net> wrote in message  4 news:EMGdne7Q9pZCzyjZnZ2dnUVZ_ridnZ2d@comcast.com... > alyciac@aol.com wrote: > F >> I'm running a VMS 6.1 system.  Last weekend, I had a batch job stopJ >> after trying and failing to delete 24 files.  (This is not a protectionI >> issue - the files just were not there to be deleted)  There is nothing J >> in my error trapping that calls for such an action.  Is there some sortG >> of maximum number of consecutive errors, after which the system will  >> terminate a job?  >   0 No there isn't a number of errors before "quit".' Be careful with a ON ERROR THEN GOTO... L in the error trap, you have to issue this ON ERROR command again before you K go back to your processing (after doing something to deal with the error -  L writing a message to the user possibly) or else it will exit, as you cannot J hit an error again, while still in you error trap (or what DCL thinks you ) are). DCL would exit in such a condition.     J > Not that I know of.  The default is to exit on the first error.  If you I > want to avoid the errors, you could test for the existance of the file   > before trying to delete it.  > J > Or, if the nonexistance of the file is an error, you can trap the error  > with ON ERROR THEN .... & > and try to do something intelligent. > H > If you want to ignore the error, you can either SET NOON or code your M > error handling to ignore the RMS-E-FNF; you'll have to grub around to find  D > the numeric code for the message, check $STATUS for that code, and  ) Instead of finding the numeric value, use L $ if f$mess($status,"ident,severity,facility") .eqs. "%DELETE-W-SEARCHFAIL"  then (whatever)   L > resume processing.  If you want to get really fancy, you can mask off the G > low order three bits of $STATUS so as to trap F, E, and W severities.   " Again, f$mess($status, "severity")   > K > DCL allows you to do just about anything you want if you want to work at   > it hard enough.  >  > <snip>   Syltrem    ------------------------------    Date: 12 Jul 2006 13:52:21 -0700) From: "Bob Gezelter" <gezelter@rlgsc.com> A Subject: Re: Number of consecutive failures allowed in batch job? A Message-ID: <1152737541.037691.46350@75g2000cwc.googlegroups.com>    alyciac,  @ It is common for a utility to issue multiple error messages, yetC actually only signal an error to the outer level when it completes.   D Without the log (and the command files) it is difficult to speculate what precisely happened.  $ - Bob Gezelter, http://www.rlgsc.com   ------------------------------    Date: 12 Jul 2006 13:57:48 -0700- From: "Doug Phillips" <dphill46@netscape.net> A Subject: Re: Number of consecutive failures allowed in batch job? A Message-ID: <1152737868.587170.82660@75g2000cwc.googlegroups.com>    alyciac@aol.com wrote:E > I'm running a VMS 6.1 system.  Last weekend, I had a batch job stop I > after trying and failing to delete 24 files.  (This is not a protection H > issue - the files just were not there to be deleted)  There is nothingI > in my error trapping that calls for such an action.  Is there some sort F > of maximum number of consecutive errors, after which the system willH > terminate a job?  The batch job in question is a series of DCL commandF > procedures which execute one after another.  After the proc with theG > deletes failed, the 'parent' proc continued with its cleanup and sent A > mail announcing the failure, so the process didn't simply stop.  >   > Any thoughts would be welcome!  > You could use the f$search lexical to verifiy the file exists.   See $HELP LEX F$SEARCH   ie.  using specific file names:  ' $ filespec = f$search("filename01.ext") , $ if filespec .nes "" then delete 'filespec'	 $! etc...    or,   E $ if f$search("filename01.ext") .nes. "" then delete filename01.exe;* 	 $! etc...    Or, with wildcards in a loop:    $d_loop:$ $ filespec = f$search("file*.tmp;*") $ if filespec .nes. "" $  then  $   delete 'filespec'  $   goto d_loop  $  endif $ exit  G In any case, using $SET NOON and/or $ON condition THEN ... in a command F file and doing your own error checking during critical steps is a good practice to follow.   F Also, unless you submit the job with /NOLOG_FILE or do a $SET NOVERIFYG at the beginning of the .com file(s), the log file should show you what  problem it's having where.   ------------------------------  % Date: Wed, 12 Jul 2006 19:41:52 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> A Subject: Re: Number of consecutive failures allowed in batch job? , Message-ID: <44B588B7.9A2C1ECE@teksavvy.com>  H If you want your batch job to continue despite any number of errors, youE need to convince the error handler to take a lunch break and not stop 
 your job :  	 $SET NOON     A If you use the ON ERROR or ON WARNING mechanisms, they are like a D catapult. They remain "loaded" until an error happens at which pointB their handling is triggered. But once triggered, the are no longer@ active and you need to re-issue an ON ERROR or ON WARNING again.  H You can isseu ON ERROR /ON WARNING multiple times and each one overrides the previous one.    ------------------------------    Date: 12 Jul 2006 17:57:02 -0700$ From: "AEF" <spamsink2001@yahoo.com>A Subject: Re: Number of consecutive failures allowed in batch job? B Message-ID: <1152752222.655784.17950@m79g2000cwm.googlegroups.com>   alyciac@aol.com wrote:E > I'm running a VMS 6.1 system.  Last weekend, I had a batch job stop I > after trying and failing to delete 24 files.  (This is not a protection H > issue - the files just were not there to be deleted)  There is nothingI > in my error trapping that calls for such an action.  Is there some sort F > of maximum number of consecutive errors, after which the system willH > terminate a job?  The batch job in question is a series of DCL commandF > procedures which execute one after another.  After the proc with theG > deletes failed, the 'parent' proc continued with its cleanup and sent A > mail announcing the failure, so the process didn't simply stop.  >   > Any thoughts would be welcome!    C It would really help a lot if you'd post an example of the problem.   G Perhaps you're counting "suberrors" as errors, the ones starting with a ! hyphen instead of a percent sign.    ------------------------------    Date: 13 Jul 2006 06:43:48 +01002 From: "Dave Weatherall" <djw-nothere@nospam.nohow>A Subject: Re: Number of consecutive failures allowed in batch job? ? Message-ID: <DTiotGxQ0bj6-pn2-93lPWRiJwvBH@dave2_os2.home.ours>   8 On Wed, 12 Jul 2006 19:23:14 UTC, alyciac@aol.com wrote:  E > I'm running a VMS 6.1 system.  Last weekend, I had a batch job stop I > after trying and failing to delete 24 files.  (This is not a protection H > issue - the files just were not there to be deleted)  There is nothingI > in my error trapping that calls for such an action.  Is there some sort F > of maximum number of consecutive errors, after which the system willH > terminate a job?  The batch job in question is a series of DCL commandF > procedures which execute one after another.  After the proc with theG > deletes failed, the 'parent' proc continued with its cleanup and sent A > mail announcing the failure, so the process didn't simply stop.  >   > Any thoughts would be welcome! >   F I usually precede a command that could go wrong, and which would _not_ be fatal, with   	$ ON EROR THEN CONTINUE  > That allows the next error to be accepted and processing will F continue. There are dangers of course; so remembering to re-establish 3 error trapping with SET NOON can then be necessary.    --   Cheers - Dave W.   ------------------------------    Date: 12 Jul 2006 15:31:41 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) % Subject: Re: OpenVMS V8.3 is in sight 3 Message-ID: <G4JssRmNfdVJ@eisner.encompasserve.org>   e In article <1152712790.343017.45010@m79g2000cwm.googlegroups.com>, "Ian Miller" <ijm@uk2.net> writes: H > there is no comparision in terms of features but I just wondered about4 > how often new versions of MS Windows are released. >   G    Well, Billy is now saying Vista in January.  Must have put a wrinkle     in Micky's Christmas plans.  J    XP shipped somewhere around late 2001?   That would be over five years.   ------------------------------    Date: 12 Jul 2006 14:33:29 -0700- From: "Doug Phillips" <dphill46@netscape.net> % Subject: Re: OpenVMS V8.3 is in sight C Message-ID: <1152740009.394446.310330@i42g2000cwa.googlegroups.com>    Ian Miller wrote: H > there is no comparision in terms of features but I just wondered about4 > how often new versions of MS Windows are released.  D Please don't think I'm defending m$, here, but Windows Service PacksC are usually significant releases (ie. WinXP, WinXP SP1, WinXP SP2). A Windows new named versions are *radical* changes to the operating D system (ie. Win98, WinME, WinXP; WinNT, Win2000, Win2003). And, m$'sC huge bundle of (what we call) layered products (Orifice, Exchanged, F Outleak, Internet Exposer, Media Splayer, etc) must be "upgraded" too,G before the release --- and often, it's one or more LP that holds things  up.   F OpenVMS releases and Windows releases are difficult to compare becauseD many Windows "improvements" are made for purely marketing reasons --G unlike OpenVMS where marketing doesn't appear to be a motivating factor + and changes *are* meant to be improvements.    ------------------------------    Date: 12 Jul 2006 14:47:13 -0700 From: "Joerg" <spi@equicon.de>$ Subject: Re: Oracle10g on IA64 8.2-1C Message-ID: <1152740833.412550.262950@b28g2000cwb.googlegroups.com>    Thans=B4ks to all!G This are not good news but news and I know that I'm not to dumb to find 
 the download.    Regards, Joerg   " Peter 'EPLAN' LANGSTOEGER schrieb:  L > In article <1152601484.410383.162330@h48g2000cwc.googlegroups.com>, "Joer= g" <spi@equicon.de> writes: @ > >I want test Oracle10g on IA64. How can I donwload a Oracle10g > >installation kit? > L > To download 10GR1 for OpenVMS Alpha you use (after accepting the license = terms) > L > http://www.oracle.com/technology/software/products/database/oracle10g/htd= ocs/openvmssoft.html > D > NOTE: You need an OTN (Oracle Technology Network) account for this > K > But to download 10G for OpenVMS I64 (which will be only 10GR2 and up) you L > have to wait for 10GR2 for OpenVMS (Alpha&I64) =3D> a couple of months lo= nger.  > L > Too bad, that Oracle for OpenVMS comes many months/years after other plat= forms.B > I remember of an (not so long ago) Oracle promise ("90 days")... >  > Good luck  >  > -- > Peter "EPLAN" LANGSTOEGER ' > Network and OpenVMS system specialist  > E-mail  peter@langstoeger.atH > A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  % Date: Wed, 12 Jul 2006 22:55:39 -0400 3 From: "Peter Weaver" <newsonly@weaverconsulting.ca> " Subject: Re: VAX 66x0 & CHARON VAX( Message-ID: <e94cna$5oh$1@emma.aioe.org>  / "Tom Linden" <tom@kednos.com> wrote in message  # news:op.tck1ztqczgicya@hyrrokkin... : > Following a recent thread on CHARON VAX I was curious if? > any such systems still running.  I can't imagine SRI actually A > going through the effort without adequate market opportunities.  >  > Tom   @ I didn't see the other thread on CHARON-VAX (Sympatico recently D stopped having a newsfeed and I have been playing around with a few > other newsfeeds trying to find a good one), but I have sold a C CHARON-VAX license to a customer who was running a 6610. The power  G supply failed the day I came to demo CHARON-VAX and they were having a  G hard time finding a replacement part on e-Bay. Since they were down to  E a single user on this system they were able to replace the 6610 with  F CHARON-VAX/XM. The license paid for itself rather quickly considering C the power demands of a 6610 with RA82 drives and the user tells me  @ that performance is much better than it was on the hardware VAX.   --   Peter Weaver www.weaverconsulting.ca  CHARON-VAX  CHARON-AXP     ------------------------------  % Date: Wed, 12 Jul 2006 23:41:06 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> " Subject: Re: VAX 66x0 & CHARON VAX, Message-ID: <44B5C0BB.84391CDE@teksavvy.com>   Peter Weaver wrote: ? > other newsfeeds trying to find a good one), but I have sold a \ > CHARON-VAX license to a customer who was running a 6610.  <vertues of Charon Vax ommitted>  B Aren't you supposed to have some "[Shameless Plug alert]" in thereH somewhere, or is that copyright of another Charon distributor ?  :-) ;-) :-) :-) :-) :-) :-) :-)    ------------------------------  # Date: Wed, 12 Jul 2006 18:01:55 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>- Subject: Re: Where Are The System Directories / Message-ID: <nOatg.402$ZP6.81@news.cpqcorp.net>    Chris L wrote:  . > What is the standard file directory system?       SYS$SYSROOT:[*...]   F    OpenVMS uses a rooted system disk structure and various searchlist G logical names, with a system root containing system-specific files for  G those systems that boot off the particular system disk and one or more  D system roots present on most every system disk, and with the common 4 system files stored over in the cluster common root.  I    Do realize that there are directory aliases used in this area, as the  D main store of files is located in a common area (SYS$COMMON:[*...], G specifically), and there are what Unix users might consider hard links  E -- what OpenVMS users call directory alias entries -- connecting the  D common directories into the system-specific system root directories.  1 > For example, where are the password files kept?   I    The OpenVMS authorization database is spread over various files, with  E the files containing the hashed passwords -- OpenVMS itself does not  I store cleartext passwords in the authorization database -- are stored in  I the file SYSUAF.DAT.  SYSUAF.DAT is located, by default, in SYS$SYSTEM:,  ? but relocation into other directories or onto other devices is  C documented and supported, and redirection is established using the   SYSUAF logical name.  G    OpenVMS hashes the password (using various user-specific strings in  B addition to the password itself) via the sys$hash_password system G service, generating (via a purdy polynomial) a quadword value, and the  I quadword value is then compared against the saved hashed quadword value;    against the saved password hash.  =    The OpenVMS User's Guide and the OpenVMS System Manager's  G (Essentials) are a good starting point, and these are available at the  D OpenVMS documentation web site <http://www.hp.com/go/openvms/doc/>. B There are various tutorials, guides and books on this and various G related topics, as well, and many of these resources are referenced in  7 the OpenVMS FAQ at <http://www.hp.com/go/openvms/faq/>.   H    Information on relocating the SYSUAF file and the other core OpenVMS D files is contained in the comments in the example command procedure D SYS$MANAGER:SYLOGICALS.TEMPLATE, as found on OpenVMS V7.2 and later.  I    What might you be up to here?  (One- or two-line questions often make  B it surprisingly difficult for me or for other folks to answer the I intended question.  The question can go in so many different directions,  H of course.  Some background might be useful here, if the above does not  address the question.)   ------------------------------    Date: 12 Jul 2006 15:33:13 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) - Subject: Re: Where Are The System Directories 3 Message-ID: <BJypu8bQk1cp@eisner.encompasserve.org>   j In article <1152721928.456102.137640@m73g2000cwd.googlegroups.com>, "Chris L" <clusardi2k@aol.com> writes:H > What is the standard file directory system? For example, where are the > password files kept?  2    Somewhere even a script kiddie could find them.  5    What, you can't find a script?  I'm not surprized.    ------------------------------   End of INFO-VAX 2006.386 ************************