[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20040205: McIDAS 2003 Environmental Variables

>From: Jim Koermer <address@hidden>
>Organization: Plymouth State University
>Keywords: 200402051340.i15DeQp1025153 McIDAS FreeBSD 5.2 environment varialbes

Hi Jim,

>In Seattle, you got McIDAS 2003 working for me on one of our new servers 
>(storm). I haven't changed anything on the system, but I'm not having 
>any success. As I remember, you had to add two McIDAS environmental 
>variables to what I had already set. I was checking the Unidata McIDAS 
>web pages, but couldn't find anything that jumped out at me on these 
>variables.
>
>Here is what I have set for user, "jim":
>
>MCHOME=/usr/local/mcidas
>MCGUI=/usr/local/mcidas/bin
>MCDATA=/home/jim/mcidas/data
>MCTABLE_WRITE=/home/jim/mcidas/data/MCTABLE.TXT
>MCTABLE_READ=/home/jim/mcidas/data/MCTABLE_TXT;/usr/local/mcidas/data/ADDESITE
> .TXT

There appears to be a typo in MCTABLE_READ.  It should be:

MCTABLE_READ=/home/jim/mcidas/data/MCTABLE.TXT;/usr/local/mcidas/data/ADDESITE.TXT

>MCPATH=/home/jim/mcidas/data:/usr/local/mcidas/data:/data/mcidas:/usr/local/mc
> idas/help
>LD_LIBRARY_PATH=/usr/local/mcidas/lib:/usr/local/mcidas/tcl:/usr/local/mcidas/
> tcl/lib
>TK_LIBRARY=/usr/local/mcidas/tcl:/usr/local/mcidas/tcl/lib
>TCL_LIBRARY=/usr/local/mcidas/tcl:/usr/local/mcidas/tcl/lib

The T*_LIBRARY definitions are probably not needed.

>MCCOMPRESS=TRUE
>
>Does this list look complete?

Outside of the typo in McTABLE_READ, the above looks OK.  The one
variable that you do not include in your list, however, is PATH.  After
logging on, what is PATH?  Does it contain $MCGUI?  Is it being reset
each time .cshrc is sourced?

>I've just about finished configuring our other new server to be our 
>primary web server and I just need to get McIDAS going to wrap it up.

If you would remind me of the login information, I could take a look
around to see what might be amiss.  Right off and without knowing the
symptoms of what is not working, I would have to guess that the problem
is somehow related to either MCTABLE_READ or the environment variable
definitions in .cshrc.  I say the latter based on the assumption that
your SHELL is csh/tcsh.  In this case, the entries in .cshrc need to
have an 'if' clause that will prevent the environment variables from
being redefined when McIDAS starts.  McIDAS is basically started when
mcenv runs to create the shared memory segments needed by McIDAS
programs.  If MCPATH gets redefined on mcenv startup, then the location
of the temporary directory created for that session, ~/.mcenv/nnnnnn
(where nnnnnn is the shared memory segment handle), is lost and things
will not work correctly.

>BTW, our BSD 5.2 system (storm) has been rock solid with LDM not 
>running--no crashes. The other new system (typhoon) is running BSD 4.9 
>and runs fine with the same LDM configuration that crashes storm. There 
>are still definite problems relating to heavy I/O with the BSD 5.x release.

Yes, I have to agree with Ted W.  The LDM is probably stressing 5.2 in
ways that it does not like.  Hopefully, 5.3 will solve some of the
issues that 5.2 created.

By the way, I won't be in today, so if you want me to poke around it
will have to be tomorrow.

Cheers,

Tom
--
NOTE: All email exchanges with Unidata User Support are recorded in the
Unidata inquiry tracking system and then made publically available
through the web.  If you do not want to have your interactions made
available in this way, you must let us know in each email you send to us.

>From address@hidden Thu Feb  5 12:52:27 2004

Hi Tom,

re: There appears to be a typo in MCTABLE_READ.  It should be:
 
MCTABLE_READ=/home/jim/mcidas/data/MCTABLE.TXT;/usr/local/mcidas/data/ADDESITE.TXT
>

Thanks, there should also be a colon instead of a semicolon. I was 
defining it properly in one place, but not in the place that was passing 
the variables to the script.

Jim