| Author |
 Topic  |
|
|
aawineco
Junior Member
USA
31 Posts |
 Posted - Jul 09 2008 : 09:48:10 AM
|
Hi,
I am running (or attempting to run) a third level anaylsis in FSL. The job produces the folders and begins to run; however, it fails without producing an error report in the cluster log. The fsl log does provide some information though. It says that the directory for the cope images does not exist. At first this was perplexing because I added a line of code to cd to those directories, and my script did so without a problem. I later noticed that the mount address listed for the directory when I cd'ed to it was different from the mount address put in by the sed substitution. It seems, based on this, that perhaps it is mounting my experiment twice, and as a result, is failing to find the correct directory. Does anyone have any suggestions for how to fix this? Below is a snip of the log illustrating what I am talking about.
/usr/local/fsl/bin/fsl_sub -T 1 -l logs -N feat5_stop -j 8225 /usr/local/fsl/bin/feat /mnt/users.q.shares/winecoff.JnA1IKQ8t1lP5jZP14jm/Analysis/FSL/lvl3cue_708/lvl3cue_cope1.gfeat/cope1.feat/design.fsf -D /mnt/users.q.shares/winecoff.JnA1IKQ8t1lP5jZP14jm/Analysis/FSL/lvl3cue_708/lvl3cue_cope1.gfeat -stop
Higher-level input files preparation
/usr/local/fsl/bin/featregapply /mnt/users.q.shares/winecoff.LgfGkfrBQ1sURnFNMjXF/Analysis/FSL/34537/34537_lvl2_cue_708.gfeat/cope1.feat
couldn't change working directory to "/mnt/users.q.shares/winecoff.LgfGkfrBQ1sURnFNMjXF/Analysis/FSL/34537/34537_lvl2_cue_708.gfeat/cope1.feat": no such file or directory
while executing
"cd [ lindex $argv 0 ]"
(file "/usr/local/fsl/bin/featregapply" line 120) |
|
|
josh.bizzell
BIAC Staff
USA
118 Posts |
Posted - Jul 09 2008 : 10:06:06 AM
|
Amy,
Can you post the cluster script?
Josh |
 |
|
|
aawineco
Junior Member
USA
31 Posts |
Posted - Jul 09 2008 : 10:32:29 AM
|
Here is the user script part. Sorry there are so many commented out parts.
# -- BEGIN USER DIRECTIVE --
# Send notifications to the following address
#$ -M winecoff@biac.duke.edu
# -- END USER DIRECTIVE --
# -- BEGIN USER SCRIPT --
# User script goes here
# NOTE:Set OUTDIR HERE. Don't modify/remove/comment-out the value in POST-USER section,
# It won't override the value you set but is a safe default in case you don't set one
OUTDIR=$EXPERIMENT/Analysis/Cluster/Job_logs
# NOTE:Set variables the usual way (the variable constructs in PRE/POST USER sections
# do special things that you don't need
#SUBJ=$1
#data location and other variables
FSLOUTDIR=${EXPERIMENT}/Analysis/FSL/lvl3cue_708
#MAINDIR=$FSLDATADIR2/${SUBJ}_run${RUN}_noMF.feat
MAINOUTPUT=${FSLOUTDIR}/lvl3cue_cope1
mkdir -p $MAINOUTPUT
OUTPUT=${MAINOUTPUT}
#FSLDATADIR=${MAINDIR}
TEMPLATEDIR=${EXPERIMENT}/Analysis/FSL/AnalysisTemplates
echo $OUTPUT
INPUT_DIR=${EXPERIMENT}/Analysis/FSL/34537/34537_lvl2_cue_708.gfeat/cope1.feat
cd ${INPUT_DIR}
echo ${INPUT_DIR}
#run FEAT on each run
INPUT1=${EXPERIMENT}/Analysis/FSL/34537/34537_lvl2_cue_708.gfeat/cope1.feat/stats/cope1.nii.gz
INPUT2=${EXPERIMENT}/Analysis/FSL/34556/34556_lvl2_cue_708.gfeat/cope1.feat/stats/cope1.nii.gz
#INPUT3=${EXPERIMENT}/Data/FSL2/20060217_45625/20060217_45625_feat_maskfix/20060217_45625_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUT4=${EXPERIMENT}/Data/FSL2/20060217_45628/20060217_45628_feat_maskfix/20060217_45628_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUT5=${EXPERIMENT}/Data/FSL2/20060220_45630/20060220_45630_feat_maskfix/20060220_45630_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUT6=${EXPERIMENT}/Data/FSL2/20060222_45642/20060222_45642_feat_maskfix/20060222_45642_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUT7=${EXPERIMENT}/Data/FSL2/20060222_45644/20060222_45644_feat_maskfix/20060222_45644_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUT8=${EXPERIMENT}/Data/FSL2/20060224_45654/20060224_45654_feat_maskfix/20060224_45654_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUT9=${EXPERIMENT}/Data/FSL2/20060227_45658/20060227_45658_feat_maskfix/20060227_45658_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTa=${EXPERIMENT}/Data/FSL2/20060302_45665/20060302_45665_feat_maskfix/20060302_45665_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTb=${EXPERIMENT}/Data/FSL2/20060302_45666/20060302_45666_feat_maskfix/20060302_45666_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTc=${EXPERIMENT}/Data/FSL2/20060306_45672/20060306_45672_feat_maskfix/20060306_45672_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTd=${EXPERIMENT}/Data/FSL2/20060307_45678/20060307_45678_feat_maskfix/20060307_45678_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTe=${EXPERIMENT}/Data/FSL2/20060310_45692/20060310_45692_feat_maskfix/20060310_45692_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTf=${EXPERIMENT}/Data/FSL2/20060313_45697/20060313_45697_feat_maskfix/20060313_45697_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTg=${EXPERIMENT}/Data/FSL2/20060320_45714/20060320_45714_feat_maskfix/20060320_45714_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTh=${EXPERIMENT}/Data/FSL2/20060322_45723/20060322_45723_feat_maskfix/20060322_45723_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTi=${EXPERIMENT}/Data/FSL2/20060323_45730/20060323_45730_feat_maskfix/20060323_45730_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTj=${EXPERIMENT}/Data/FSL2/20060404_45756/20060404_45756_feat_maskfix/20060404_45756_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTk=${EXPERIMENT}/Data/FSL2/20060421_45809/20060421_45809_feat_maskfix/20060421_45809_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTl=${EXPERIMENT}/Data/FSL2/20060427_45821/20060427_45821_feat_maskfix/20060427_45821_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTm=${EXPERIMENT}/Data/FSL2/20060428_45825/20060428_45825_feat_maskfix/20060428_45825_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTn=${EXPERIMENT}/Data/FSL2/20060428_45829/20060428_45829_feat_maskfix/20060428_45829_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTo=${EXPERIMENT}/Data/FSL2/20060502_45835/20060502_45835_feat_maskfix/20060502_45835_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTp=${EXPERIMENT}/Data/FSL2/20060503_45837/20060503_45837_feat_maskfix/20060503_45837_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTq=${EXPERIMENT}/Data/FSL2/20060511_45860/20060511_45860_feat_maskfix/20060511_45860_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUTr=${EXPERIMENT}/Data/FSL2/20060518_45869/20060518_45869_feat_maskfix/20060518_45869_lvl2_MF.gfeat/cope11.feat/stats/cope1.nii.gz
#INPUT2=${FSLDATADIR2}/${SUBJ}_run2_MF.feat
#INPUT3=${FSLDATADIR2}/${SUBJ}_run3_MF.feat
#INPUT4=${FSLDATADIR2}/${SUBJ}_run4_MF.feat
#INPUT5=${FSLDATADIR2}/${SUBJ}_run5_MF.feat
#INPUT6=${FSLDATADIR2}/${SUBJ}_run6_MF.feat
echo $INPUT1
echo $INPUT2
cd ${TEMPLATEDIR}
for i in 'lvl3cue_template.fsf'; do
sed -e 's@OUTPUT@'$OUTPUT'@g'\
-e 's@INPUT2@'$INPUT2'@g'\
-e 's@INPUT1@'$INPUT1'@g'\ <$i> ${FSLOUTDIR}/lvl3cue_cope1_708.fsf
done
cd ${MAINOUTPUT}
#run the newly created fsf files
feat ${FSLOUTDIR}/lvl3cue_cope1_708.fsf
# -- END USER SCRIPT -- # |
|
|
|
josh.bizzell
BIAC Staff
USA
118 Posts |
Posted - Jul 09 2008 : 11:19:11 AM
|
Unfortunately, the script itself looks fine.
It shouldn't be mounting the experiment twice, unless you have two "biacmount" commands in the script.
Josh |
|
|
|
aawineco
Junior Member
USA
31 Posts |
Posted - Jul 09 2008 : 12:44:13 PM
|
| Is there any other reason you can think of that would cause the addresses to be different? |
|
|
|
aawineco
Junior Member
USA
31 Posts |
Posted - Jul 09 2008 : 12:48:14 PM
|
| Nevermind. I figured it out. |
|
|
| |
Topic |
|
FSL failure: re-mounting experiment
