modis_GEO.csh: ERROR: Unable to determine start time from L1A file.

By Thomas.Browning Date 2011-05-26 13:19

Hi,

I have tried to create a GEO file using a L1A MODIS image using the seaDAS GUI.
An error occurs, which is documented in the terminal as:

seadas@seadas-vm:/disk01/seadas/seadas/Test$ seadas -em
IDL Version 7.0 (linux x86 m32). (c) 2007, ITT Visual Information Solutions

% Embedded IDL: NASA GSFC SeaDAS Development, SeaDAS.
% Embedded IDL: NASA GSFC SeaDAS Development, SeaDAS.
SeaDAS Version 6.2 Update 2 (pid = 3440)
unix cmd =
$SEADAS/run/scripts/modis_GEO.csh /disk01/seadas/seadas/Test/A2010299125000.L1A_LAC.x.hdf -o /disk01/seadas/seadas/Test/A2010299125000.GEO -geocheck_threshold 95
/disk01/seadas/seadas/run/bin/modis_timestamp: Exec format error. Binary file not executable.
modis_GEO.csh: ERROR: Unable to determine start time from L1A file.

exit_status= 1

I am using windows 7 (64bit) with the standard seaDAS Ubuntu/VMware player set up outlined on the seaDAS installation page (http://seadas.gsfc.nasa.gov/seadasva.html). I have also downloaded and extracted the following additional processing files (my aim is to only process MODIS aqua images):

seadas_dem_modis.gz.tar
seadas_processing_linux_64.gz.tar
seadas_benchmarks.gz.tar
seadas_modisa.gz.tar
seadas_processing_common.gz.tar

I have seen some other fairly recent posts on the forum which relate to an issue with modis_timestamp on 64bit systems, but they do not seem to really apply to my L1A processing aim...

Any help would be great.

Cheers,
Tom

By gwyn Date 2011-05-26 14:25

Tom,

The VM is set up as a 32-bit machine, independent of your 64-bit host system. You should have installed executables from seadas_processing_linux.tar.gz instead of seadas_processing_linux_64.tar.gz .

If the problem is still not resolved, please enter the following at the command line and post the results:
uname -a ls -l /disk01/seadas/seadas/run/bin/modis_timestamp file /disk01/seadas/seadas/run/bin/modis_timestamp
Gwyn

By Thomas.Browning Date 2011-05-27 12:27

Gwyn, thanks for your response - I reinstalled seadas with seadas_processing_linux.tar.gz instead of seadas_processing_linux_64.tar.gz. I still get a 'start time from L1A file' problem (although doing the former will help avoid future issues im sure!)...

SeaDAS Version 6.2 Update 2 (pid = 3309)
unix cmd =
$SEADAS/run/scripts/modis_GEO.csh /disk01/seadas/seadas/Test/A2010299125000.L1A_LAC.x.hdf.bz2 -o /disk01/seadas/seadas/Test/A2010299125000.GEO -geocheck_threshold 95
*** /home/seadas/seadas/run/bin3/ncdump: ncopen failed on /disk01/seadas/seadas/Test/A2010299125000.L1A_LAC.x.hdf.bz2�i
modis_GEO.csh: ERROR: Unable to determine start time from L1A file.

exit_status= 1

I then entered the following as instructed.....

seadas@seadas-vm:/disk01/seadas/seadas/Test$ uname -a
Linux seadas-vm 2.6.32-26-generic #48-Ubuntu SMP Wed Nov 24 09:00:03 UTC 2010 i686 GNU/Linux
seadas@seadas-vm:/disk01/seadas/seadas/Test$ ls -l /disk01/seadas/seadas/run/bin/modis_timestamp
-rwxr-xr-x 1 seadas seadas 1107746 2011-04-06 12:54 /disk01/seadas/seadas/run/bin/modis_timestamp
seadas@seadas-vm:/disk01/seadas/seadas/Test$ file /disk01/seadas/seadas/run/bin/modis_timestamp
/disk01/seadas/seadas/run/bin/modis_timestamp: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), statically linked, for GNU/Linux 2.6.9, not stripped

Any ideas?

Cheers,
Tom

By gwyn Date 2011-05-27 12:45

Tom,

You need to unpack your file!

bunzip2 A2010299125000.L1A_LAC.x.hdf.bz2

Then run with A2010299125000.L1A_LAC.x.hdf

By Thomas.Browning Date 2011-05-27 13:10

Gwyn, of course....forgot to do that after getting the files again. Anyway - seem to get same output....

SeaDAS Version 6.2 Update 2 (pid = 3600)
unix cmd =
$SEADAS/run/scripts/modis_GEO.csh /disk01/seadas/seadas/Test/A2010299125000.L1A_LAC.x.hdf -o /disk01/seadas/seadas/Test/A2010299125000.GEO -geocheck_threshold 95
*** /home/seadas/seadas/run/bin3/ncdump: ncopen failed on /disk01/seadas/seadas/Test/A2010299125000.L1A_LAC.x.hd����i
modis_GEO.csh: ERROR: Unable to determine start time from L1A file.

exit_status= 1

Thanks for your help on this!
Tom

By @sean Date 2011-05-27 15:22

Tom,

The problem is that ncdump is trying to open: /disk01/seadas/seadas/Test/A2010299125000.L1A_LAC.x.hd����i
and not /disk01/seadas/seadas/Test/A2010299125000.L1A_LAC.x.hdf

You are not the first to run into this problem.
http://oceancolor.gsfc.nasa.gov/forum/oceancolor/topic_show.pl?pid=17065#pid17065
http://oceancolor.gsfc.nasa.gov/forum/oceancolor/topic_show.pl?pid=16978#pid16978

Unfortunately, I have not had time to see if I could replicate the problem and come up with a solution.
It seems that running the command from the directory which contains the files works for some. This
suggests the issue is with how the filepath is built (the modis_GEO.csh script splits the path from the filename
for some funtions, then rebuilds the full path to run the executables).

You could also try the python versions of the script that was delivered with 6.2 Update 2:
$SEADAS/run/scripts/modis_GEO.py

Sean

***********************************************************
$SEADAS/run/scripts/modis_GEO.py --help
usage:
modis_GEO.py [OPTIONS] MODIS_L1A_file
or
modis_GEO.py --par parameter_file [OPTIONS]

options:
--version    show program's version number and exit
-h, --help show this help message and exit
-p PARFILE, --parfile=PARFILE
Parameter file containing program inputs
-o GEOFILE, --output=GEOFILE
Output filename
-a ATT1, --att1=ATT1 Input attitude file 1 (chronological)
-A ATT2, --att2=ATT2 Input attitude file 2 (chronological)
-e EPH1, --eph1=EPH1 Input ephemeris file 1 (chronological)
-E EPH1, --eph2=EPH1 Input ephemeris file 2 (chronological)
-c THRESHOLD, --threshold=THRESHOLD
% of geo-populated pixels required to pass geocheck
validation test
-d, --enable-dem Enable MODIS terrain elevation correction
-v, --verbose    print status messages
-l, --log    Save processing log file(s)
***********************************************************

By Thomas.Browning Date 2011-05-31 11:02

Hi Sean,

Thanks for your response. I have fiddled around a bit, still dont have it working but will keep trying.

Do you think it is a problem with 64bit windows? Would installing on a windows 32bit machine or on a dual boot linux machine fix the problem (and cause less issues in general) do you think?

Thanks for your help,
Tom

By @sean Date 2011-05-31 17:40

Tom,

No, it's not likely a 64bit windows issue. The VM obviously works, and when you run it you
are effectively in a 32bit linux world...right now we're thinking it's a shell issue - or ASCII vs Unicode,
really just grasping at straws now, since we haven't been able to replicate it here...yet...

Sean

By Thomas.Browning Date 2011-06-03 11:03

Hi Sean,

Thanks for your help - I have got it working. The way I got round it was simply by shortening the L1A file name. The file name was too long for the program, and shortening it to A.L1B_LAC did the trick. The .L1B_LAC needs to stay in the file name or it does not work...

Cheers,
Tom

By @sean Date 2011-06-04 00:09

Thanks, Tom.
That's a clue for us to go on...file/path name length is involved...

Sean

By King.Jerome Date 2011-06-09 11:44

Hi,

FYI,
We have run into this same error message this last couple of days at the Cornell Remote Sensing class and path length definitely seem to be the issue.
The class uses a central hardisk with all the students' account and a couple symbolic links are used.
It seems that SeaDAS uses the full path length and therefore it looks like we are exceeding the max length SeaDAS will tolerate.
As soon as we reduced the path length by renaming some directories, the error went away.
Thanks for the tip,
Jerome.