20000322: duplicate McIDAS Mollweide images

>From: Mike Leuthold <address@hidden>
>Organization: University of Arizona
>Keywords: 200003221639.JAA04528 ldm-mcidas Mollweide lwtoa3 duplicate broken 
>pipe LA

Mike,

>Recently, I've been getting duplicate Mollweide images.  I check
>the archives for this error and there was mention of something like this
>happening when lwtoa3 was called twice due to a broken pipe.  I didn't see
>this in my logs.

I see from your second email that you did find broken pipe messages in your
ldmd.log file.

>Any other ideas?  Here is the output from the
>decoder....

This is the output from your ldm-mcidas log file.

>Mar 21 22:36:23 lwtoa3[3582935]: Starting Up
>Mar 21 22:36:23 lwtoa3[3582935]: changing to directory /u/mcidas/data
>Mar 21 22:36:23 lwtoa3[3582935]: decoding "LWTOA3 110 DIALPROD=UY    81 223601
> "
>Mar 21 22:36:23 lwtoa3[3582935]: PRODUCT CODE=UY          81          223601
>Mar 21 22:36:25 lwtoa3[3582935]:  Done -- AREA= 1100
>Mar 21 22:36:25 lwtoa3[3582935]: Exiting
>Mar 21 22:36:25 lwtoa3[3590637]: Starting Up
>Mar 21 22:36:25 lwtoa3[3590637]: changing to directory /u/mcidas/data
>Mar 21 22:36:25 lwtoa3[3590637]: decoding "LWTOA3 110 DIALPROD=UY    81 223601
> "
>Mar 21 22:36:25 lwtoa3[3590637]: PRODUCT CODE=UY          81          223601
>Mar 21 22:36:27 lwtoa3[3590637]:  Done -- AREA= 1101
>Mar 21 22:36:27 lwtoa3[3590637]: Exiting
>
>Here is the output from la
>LA 1000 1039
>  area  ss  yyddd hhmmss  lcor  ecor lr er zr  lsiz  esiz z    bands
>  ---- ---- ----- ------ ----- ----- -- -- -- ----- ----- -    -----
>  1000  70  ***** 210000  9826  9681  1  1  1   350   640 1 ...4............
>  1001  70  ***** 210000  9826  9681  1  1  1   350   640 1 ...4............
>  1002  70  *****      0  9826  9681  1  1  1   350   640 1 ...4............
>  1003  70  *****      0  9826  9681  1  1  1   350   640 1 ...4............
>  1004  70  *****  30000  9826  9681  1  1  1   350   640 1 ...4............
>  1005  70  *****  60000  9826  9681  1  1  1   350   640 1 ...4............
>  1006  70  *****  90000  9826  9681  1  1  1   350   640 1 ...4............
>  1007  70  *****  90000  9826  9681  1  1  1   350   640 1 ...4............
>  1008  70  ***** 120000  9826  9681  1  1  1   350   640 1 ...4............
>  1009  70  ***** 150000  9826  9681  1  1  1   350   640 1 ...4............
>  1010  70  ***** 150000  9826  9681  1  1  1   350   640 1 ...4............

By the way.  LA is one of those programs that are goint to go away.  The
overflow in the format for the yyddd field is indicative of the date being
in the yyyddd format.  I made an addendum to McIDAS that contained an
updated version of LA's source file, la.pgm, so you might want to grab
it if anything you are doing (e.g., ROUTE PostProcess BATCH invocations)
depends on extracting valid fields from LA output.

>I'm going to test things further.
>thanks.
>
>>From address@hidden  Wed Mar 22 10:09:34 2000
>Subject: duplicate McIDAS Mollweide images (fwd)
>
>
>Tom, 
>I take it back about not seeing a broken pipe.  The logs do show it.  Now
>to find out why.....

I havn't had the chance to really investigate why the first invocation
of lwtoa3 ends before the LDM product.  It is likely that there is some
extra stuff at the end of the product that lwtoa3 doesn't eat up before
it exits.  The LDM will see that it still has some product that should
be sent to lwtoa3 when lwtoa3 goes away.  This will be interpreted as
lwtoa3 terminating early, and a second invocation will be run.  The
fact that there is only one duplicate reflects the fact that the LDM
only retries the lwtoa3 invocation once.

If you find anything out with respect to the early exiting of lwtoa3,
please send it along.

Tom