[daip] ZSTRTA.EXE

[Eric Greisen]

Carlos De Breuck writes:

 > Thanks for your suggestions. You seem to have narrowed down the problem,
 > but I haven't been able to solve it yet.

    These are most mysterious.

 > I tried un-locking the SPD000000; file as you suggested, without succes.
 > Next, I tried to rebuild the whole directory for one machine (called
 > giscours, but the problem is the same for all of them), using SYSETUP.

     This says that it is not a single file locked unless it is one
shared by the machines.  The PW file (see below) is such a file.

 > This gets me it bit further, but gives the following errors (using notv):
 > 
 > TVDEVS.SH: Starting TPMON daemons on GISCOURS asynchronously...
 > Starting up 31DEC00 AIPS with normal priority
 > ZDCHI1: ZERROR: ON FILE DA00:SPD000000;
 > ZDCHI1: ZERROR: IN ZDAOPN ERRNO = 13 (Permission denied)
 >  ZDCHIN: COULD NOT READ PARAMETER FILE
 >  ZDCHIN: (USING MINIMUM SYSTEM CONFIGURATION)
 > Begin the one true AIPS number 1 (release of 31DEC00) at priority =   0 
 > DADEVS.PL: This program is untested under Perl version 5.006
 > AIPS 1: ZERROR: ON FILE DA00:SPD000000;
 > AIPS 1: ZERROR: IN ZDAOPN ERRNO = 13 (Permission denied)
 >  ZDCHIN: COULD NOT READ PARAMETER FILE
 >  ZDCHIN: (USING MINIMUM SYSTEM CONFIGURATION)
 > AIPS 1: You are NOT assigned a TV device or server
 > AIPS 1: You are NOT assigned a graphics device or server
 > AIPS 1: Enter user ID number
 > ?134
 > 
 > (and then it hangs on process AIPS1)
 > 
 > This seems like a simple permission problem, so I changed all the files in 
 > $NET0/GISCOURS to writeable for all users:

    Yes but it tells us something - the AIPS.EXE actually came up and
talked to you and read the number.  The next thing it does is check
the PW file.  I note that that file is linked to some other machine's
DA00 area.  Perhaps it is the one locked.

 > 
 > -rw-rw-rw-   1 aips     aips      102400 Mar  7 15:45 ACD000000;
 > lrwxrwxrwx   1 aips     aips          44 Mar  7 15:45 GRD000000; ->
 > /home/soft/osf/aips/DA00/TARIQUET/GRD000000;
 > -rw-rw-rw-   1 aips     aips      561152 Mar  7 15:45 MED000001;
 > -rw-rw-rw-   1 aips     aips      561152 Mar  7 15:45 MED000002;
 > -rw-rw-rw-   1 aips     aips      561152 Mar  7 15:45 MED000003;
 > -rw-rw-rw-   1 aips     aips      196608 Mar  7 15:45 MED000004;
 > -rw-rw-rw-   1 aips     aips      561152 Mar  7 15:45 MED000005;
 > -rw-rw-rw-   1 aips     aips      561152 Mar  7 15:45 MED000006;
 > -rw-rw-rw-   1 aips     aips      561152 Mar  7 15:45 MED000007;
 > -rw-rw-rw-   1 aips     aips      561152 Mar  7 15:45 MED000008;
 > lrwxrwxrwx   1 aips     aips          44 Mar  7 15:45 PWD000000; ->
 > /home/soft/osf/aips/DA00/TARIQUET/PWD000000;
 > -rw-rw-rw-   1 aips     aips        1024 Mar  7 15:45 SPD000000;
 > -rw-rw-rw-   1 aips     aips      165888 Mar  7 15:45 TCD000001;
 > -rw-rw-rw-   1 aips     aips       62464 Mar  7 15:45 TDD000004;
 > -rw-rw-rw-   1 aips     aips        1024 Mar  7 15:45 TPD001001;
 > -rw-rw-rw-   1 aips     aips        1024 Mar  7 15:45 TPD001002;
 > -rw-rw-rw-   1 aips     aips        1024 Mar  7 15:45 TPD001003;
 > -rw-rw-rw-   1 aips     aips        1024 Mar  7 15:45 TPD001004;
 > -rw-rw-rw-   1 aips     aips        1024 Mar  7 15:45 TPD001005;
 > 
 > After this change, the errors disappear, but it hangs on ZSTRTA.EXE even
 > before it gets to asking the user number.
 > 
 > So, I moved to your next option, and edited $AIPGUNIX/ZSTRTA.FOR
 > to change      
 > CALL ZDCHIN (.TRUE., IBUF)
 > to 
 > CALL ZDCHIN (.FALSE., IBUF)
 > 
 > This gets me a bit further. It brings up the Copyright 2002 message (which
 > is strange given that I'm using the 31DEC00 version). It gets to this:

      We cheat and show today's date here.

 > 
 > ...
 > TVDEVS.SH: Starting TPMON daemons on GISCOURS asynchronously...
 > Starting up 31DEC00 AIPS with normal priority
 >  ----------------------------------------------------------------------  
 >     Copyright (C) 2002                                                   
 >    Associated Universities, Inc. Washington DC, USA.                     
 > ...
 >                           Charlottesville, VA 22903-2475 USA             
 >  ----------------------------------------------------------------------  
 > Begin the one true AIPS number 1 (release of 31DEC00) at priority =   0 
 > DADEVS.PL: This program is untested under Perl version 5.006
 > STARTPMON: [GISCOURS] Starting TPMON1 with output SUPPRESSED
 > STARTPMON: [GISCOURS] Starting TPMON2 with output SUPPRESSED
 > 
 > And then it hangs on AIPS1.

   Is the process AIPS1 actually running and is it getting cpu time?
This is an odd result because, when the SP file was unreadable, the
AIPS.EXE actually started.

 > 
 > I then tried to see how far it got in ZSTRTA.EXE by adding in some WRITE
 > statement. It goes almost to the end:
 > C                                       Activate AIPSx (x = MYPOPS).
 >       AIPSN = 'AIPS  '
 > 	WRITE(*,*)'It gets to here...'
 >       CALL ZACTV8 (AIPSN, MYPOPS, VERSIN, PID, IERR)
 > 	WRITE(*,*)'but not here...'
 > C                                       Never expect to get here, but
 > C                                       just in case.
 >       IF (IERR.EQ.0) GO TO 999
 >          WRITE (MSGTXT,1060) IERR
 >  990     CALL MSGWRT (8)
 > C
 > 
 > So, it seems to hang on the CALL ZACTV8.
 > 
 > I haven't been able to get further than that, not even when I run AIPS
 > from the aips account.
 > 
 > Can you give me some more suggestions?

There is in $SYSUNIX a file called WHOLOCKED.C.  You could try
compiling this program and running it on various files.  If it says a
file is locked then the system thinks it is - but the info on the
process ID may be bogus.

Are there AIPS files in /tmp ?

I am a bit at a loss.

Eric Greisen