| Author |
Topic  |
|
|
Adria
New Member

19 Posts |
Posted - Dec 14 2011 : 09:19:10 AM
|
I've been having trouble with the "randomise" command in FSL. I submitted the job using qsub and a script I wrote, but every job I submit comes back with this error: "----JOB [fslvbm1.sh.3063420] START [Tue Dec 13 11:26:43 EST 2011] on HOST [node6 0]---- randomise options: -i GM_mod_merg_s2 -m GM_mask -o testdata -d design.mat -t des ign.con -T -n 5000 -V ** ERROR (nifti_image_read): failed to find header file for 'GM_mod_merg_s2' ** ERROR: nifti_image_open(GM_mod_merg_s2): bad header info ERROR: failed to open file GM_mod_merg_s2 Loading Data: ERROR: Could not open image GM_mod_merg_s2 Image Exception : #22 :: Failed to read volume GM_mod_merg_s2 ERROR: Program failed
An exception has been thrown Failed to read volume GM_mod_merg_s2Trace: read_volume4DROI."
Why is it not able to read the header information? What can I do to make it work?
It displays with FSLview no problem, and the "randomise" command will execute when just used as a command line, but not with qsub.
Thanks!
|
|
|
syam.gadde
BIAC Staff
    
USA
421 Posts |
Posted - Dec 14 2011 : 09:43:03 AM
|
Does it work if you give it the full path? i.e. /mnt/BIAC/...
Or if you preface the randomise command with a cd to that directory (inside the qsub script)? |
 |
|
|
petty
BIAC Staff
    
USA
453 Posts |
Posted - Dec 14 2011 : 10:00:57 AM
|
| also, how large is the file ... It could be that it so large there isn't enough available memory. |
 |
|
|
Adria
New Member

19 Posts |
Posted - Dec 14 2011 : 10:15:19 AM
|
| I edited the path and it submitted. I haven't received an email yet, so I'm hoping it's actually processing the job and that was the problem. |
 |
|
|
petty
BIAC Staff
    
USA
453 Posts |
Posted - Dec 14 2011 : 10:29:16 AM
|
the job log should be writing to your home directory, so you can monitor the log there.
Also, looks like this job running randomise ... per your other post, if you export FSLSUBMIT=1 inside your submission script, and change randomise to "randomise_parallel" with the same inputs, then it will split up your processing across multiple nodes. |
 |
|
| |
Topic  |
|