 |
 |
Finally, I have reinstalled fsl, condor, reconfigure condor and the computations are running. Rene Dne 7.12.2013 20:49, René Labounek napsal(a): > Hi Michael, > > fslmats works without problems. In the same terminal, I wrote > condor_run whoami, the output was labounek (my username), so there > should not be a problem. Then I wrote in same terminal bedpostx > /path_to_data_folder and nothing again, only output in monitor 0 > slices processed. > > In .bashrc, there are these two rows about fsl software. > > . /etc/fsl/5.0/fsl.sh > source /etc/fsl/fsl.sh > > If condor would have a problem with acess to read or write data on > hard disk, condor would write it into terminal, I think. But for all > cases, data are on ntfs hard disk mounted with following parameters: > ntfs rw,exec,nosuid,nodev,allow_other,blksize=4096 0 0 > > Rene > > Dne 6.12.2013 14:53, Michael Hanke napsal(a): >> Hi, >> >> I can only confirm what has been stated already. The problem you are >> seeing is a symptom of an insuffcient configuration. FSL needs to have >> its config file sourced in the target environment (shell session). The >> Debian package does this for some parts of FSL automatically (in order >> to enable GUI integration without having to force a particular >> configuration on all users of a system). >> >> But in order for FSL to work as described in the manual, it needs to be >> configured as described in the manual -- previous messages in this >> thread point out how. >> >> When FSL submits jobs to Condor the execution environment on the target >> machine inherits the environment settings of the submitting user. If you >> can run >> >> fslmaths >> >> in the terminal from which you kick off the bedpostx analysis, things >> should just work. >> >> There is one catch (which should _not_ apply here): Condor can be set up >> in a way that executes jobs under a different user. In this unlikely >> case, more tweaking would need to be done. To verify that this is not >> the problem do: >> >> condor_run whoami >> >> This should yield your username (might take a while till the job gets >> executed, depending on the cluster/machine load). >> >> Michael >> >> >> >> On Fri, Dec 06, 2013 at 02:31:44PM +0100, Labounek René wrote: >>> Condor status writes to me that cores are unclaimed after >>> fsl5.0-bedpostx comand. >>> >>> Name OpSys Arch State Activity LoadAv Mem >>> ActvtyTime >>> >>> slot1@ LINUX X86_64 Unclaimed Idle 0.040 3981 >>> 1+22:14:52 >>> slot2@ LINUX X86_64 Unclaimed Idle 0.000 3981 >>> 1+22:15:21 >>> slot3@ LINUX X86_64 Unclaimed Idle 0.000 3981 >>> 1+22:15:22 >>> slot4@ LINUX X86_64 Unclaimed Idle 0.000 3981 >>> 1+22:15:23 >>> Total Owner Claimed Unclaimed Matched >>> Preempting Backfill >>> >>> X86_64/LINUX 4 0 0 4 0 >>> 0 0 >>> >>> Total 4 0 0 4 0 >>> 0 0 >>> >>> Maybe it is important and I forgot to write it. Between last >>> analysis when the estimation worked and when it did not work, I have >>> installed freesurfer package on my PC with this manual >>> http://simnibs.de/installation/installfsandfsl . Since that time, my >>> terminal writes next rows on each terminal start. >>> >>> -------- freesurfer-Linux-centos6_x86pub-v5.3.0 -------- >>> FREESURFER_HOME /usr/local/freesurfer >>> FSFAST_HOME /usr/local/freesurfer/fsfast >>> FSF_OUTPUT_FORMAT nii.gz >>> SUBJECTS_DIR /usr/local/freesurfer/subjects >>> MNI_DIR /usr/local/freesurfer/mni >>> labounek@PC-E229-455:~$ >>> >>> Best, >>> Rene >>> >>> Cituji "[log in to unmask]" <[log in to unmask]>: >>> >>>> Hi, >>>> >>>> Great. I don't know if monitor has been improved, but as long as I >>>> know it >>>> never worked as it was meant. You might need to check condor_status. >>>> >>>> Alexandre Manh~aes Savio <[log in to unmask]> >>>> Grupo de Inteligencia Computacional <http://www.ehu.es/ccwintco> >>>> Departamento de CCIA >>>> UPV/EHU >>>> >>>> >>>> On 4 December 2013 20:17, René Labounek >>>> <[log in to unmask]> wrote: >>>> >>>>> The errors disappeared but the computer is still not computing. After >>>>> comand: >>>>> /media/407285B97285B3E6/Data/humanconnectome/Q3/151526/T1w/Diffusion.bedpostX/monitor, >>>>> >>>>> there is written only one output: 0 slices processed. >>>>> >>>>> Rene >>>>> >>>>> Dne 4.12.2013 17:28, [log in to unmask] napsal(a): >>>>> >>>>> Hi René, >>>>> >>>>> .bash_profile doesn't exist initially. >>>>> >>>>> I see that you are running fsl using "fsl5.0-". >>>>> >>>>> Could you please execute these two commands: >>>>> >>>>> echo ". /etc/fsl/5.0/fsl.sh" >> ~/.bashrc >>>>> echo "if [ -f ~/.bashrc ]; then . ~/.bashrc; fi" >> ~/.bash_profile >>>>> >>>>> It will add in the first command the settings to your interactive >>>>> terminals and in the second anything you have .bashrc to your >>>>> remote login >>>>> shells (which are used by other softwares as well). >>>>> >>>>> After that, try the experiment again in a new terminal. >>>>> >>>>> Cheers, >>>>> Alex >>>>> >>>>> >>>>> >>>>> Alexandre Manh~aes Savio <[log in to unmask]> >>>>> Grupo de Inteligencia Computacional <http://www.ehu.es/ccwintco> >>>>> Departamento de CCIA >>>>> UPV/EHU >>>>> >>>>> >>>>> On 4 December 2013 16:53, René Labounek >>>>> <[log in to unmask]>wrote: >>>>> >>>>>> Helo Alex, >>>>>> I did not set any of these .bash files before. I have added >>>>>> neurodebian >>>>>> to my repositary and installed fsl. Then I have started to use >>>>>> bedpostx >>>>>> function but it worked only on one processor. I have found this >>>>>> instructions about paralelising with condor and installed condor. >>>>>> >>>>>> >>>>>> http://neuro.debian.net/blog/2012/2012-03-09_parallelize_fsl_with_condor.html >>>>>> >>>>>> >>>>>> In /etc/fsl/fsl.sh, I have changed: export FSLPARALLEL=condor. The >>>>>> calculation was working on all my processors. Since that time, I >>>>>> have not >>>>>> changed anything. Today, I wanted to start new calculation and it >>>>>> is not >>>>>> working. >>>>>> >>>>>> I was looking into all files you have written but nowhere I did >>>>>> not find >>>>>> any variable called FSL or something like that. Maybe problem is >>>>>> here but I >>>>>> do not know which variables I should define. If I understand >>>>>> these files >>>>>> serves for definition of global bash variables. On my PC, I have >>>>>> found >>>>>> .bashrc and /etc/bash.bashrc, not found ~/.bash_profile. >>>>>> >>>>>> In error, there is writen /bin/zeropad: not found and >>>>>> /bin/imtest: not >>>>>> found. I am not sure what this files should do but truly I have >>>>>> not them on >>>>>> hard disk. >>>>>> >>>>>> Even, I am not sure I should write this problem here or on some >>>>>> neurodebian forum. >>>>>> >>>>>> Rene >>>>>> >>>>>> >>>>>> Dne 4.12.2013 16:03, [log in to unmask] napsal(a): >>>>>> >>>>>> Hello René, >>>>>> >>>>>> Have you set your ~/.bash_profile file? >>>>>> It should source either your fsl settings or your .bashrc or >>>>>> /etc/bash.bashrc file if you have them changed for FSL. >>>>>> If that is the case, more info here: >>>>>> http://mywiki.wooledge.org/DotFiles >>>>>> >>>>>> I hope this helps. >>>>>> >>>>>> Cheers, >>>>>> Alex >>>>>> >>>>>> Alexandre Manh~aes Savio <[log in to unmask]> >>>>>> Grupo de Inteligencia Computacional <http://www.ehu.es/ccwintco> >>>>>> Departamento de CCIA >>>>>> UPV/EHU >>>>>> >>>>>> >>>>>> On 4 December 2013 15:50, René Labounek >>>>>> <[log in to unmask]>wrote: >>>>>> >>>>>>> Dear FSL users, >>>>>>> >>>>>>> I am using condor for bedpostx paralelising on Ubuntu 12.04. The >>>>>>> computation worked but suddenly it does not. Does anyone >>>>>>> have same problem? >>>>>>> Some problem with update or something else? >>>>>>> >>>>>>> Here is my folder with dMRI data. I think everything should be >>>>>>> right. >>>>>>> >>>>>>> labounek@PC-E229-455:/media/407285B97285B3E6/Data/humanconnectome/Q3/151526/T1w/Diffusion$ >>>>>>> >>>>>>> *ls* >>>>>>> bedpostx_qsub.RZG87UQEgeo23* bvals bvecs data.nii.gz >>>>>>> grad_dev.nii.gz nodif_brain_mask.nii.gz* >>>>>>> >>>>>>> Now, when I will use bedpostx comand with default parameter >>>>>>> setting I >>>>>>> will get the condor output: >>>>>>> >>>>>>> *fsl5.0-bedpostx >>>>>>> /media/407285B97285B3E6/Data/humanconnectome/Q3/151526/T1w/Diffusion* >>>>>>> >>>>>>> >>>>>>> ______________________________________________________________________________________________________________________________________ >>>>>>> >>>>>>> subjectdir is >>>>>>> /media/407285B97285B3E6/Data/humanconnectome/Q3/151526/T1w/Diffusion >>>>>>> >>>>>>> Making bedpostx directory structure >>>>>>> Queuing preprocessing stages >>>>>>> Queuing parallel processing stage >>>>>>> Queuing post processing stage >>>>>>> >>>>>>> Type >>>>>>> /media/407285B97285B3E6/Data/humanconnectome/Q3/151526/T1w/Diffusion.bedpostX/monitor >>>>>>> >>>>>>> to show progress. >>>>>>> Type >>>>>>> /media/407285B97285B3E6/Data/humanconnectome/Q3/151526/T1w/Diffusion.bedpostX/cancel >>>>>>> >>>>>>> to terminate all the queued tasks. >>>>>>> >>>>>>> You will get an email at the end of the post-processing stage. >>>>>>> >>>>>>> ______________________________________________________________________________________________________________________________________ >>>>>>> >>>>>>> >>>>>>> The Diffusion.bedpostx folder is created including tipical >>>>>>> folders and >>>>>>> files. >>>>>>> >>>>>>> >>>>>>> labounek@PC-E229-455:/media/407285B97285B3E6/Data/humanconnectome/Q2/Q2_diffusion_preprocessed/499566/T1w/Diffusion.bedpostX$ >>>>>>> >>>>>>> *ls* >>>>>>> *cancel commands.txt diff_slices logs monitor xfms* >>>>>>> >>>>>>> >>>>>>> After comand: >>>>>>> */media/407285B97285B3E6/Data/humanconnectome/Q3/151526/T1w/Diffusion.bedpostX/monitor >>>>>>> *; I will get lots error messages lokking like these shown below >>>>>>> and >>>>>>> the calculation is not working. >>>>>>> >>>>>>> __________________________________________________________________________________________________________________________________________________ >>>>>>> >>>>>>> /media/407285B97285B3E6/Data/humanconnectome/Q3/151526/T1w/Diffusion.bedpostX/monitor: >>>>>>> >>>>>>> 1: >>>>>>> /media/407285B97285B3E6/Data/humanconnectome/Q3/151526/T1w/Diffusion.bedpostX/monitor: >>>>>>> >>>>>>> /bin/zeropad: not found >>>>>>> /media/407285B97285B3E6/Data/humanconnectome/Q3/151526/T1w/Diffusion.bedpostX/monitor: >>>>>>> >>>>>>> 1: >>>>>>> /media/407285B97285B3E6/Data/humanconnectome/Q3/151526/T1w/Diffusion.bedpostX/monitor: >>>>>>> >>>>>>> /bin/imtest: not found >>>>>>> /media/407285B97285B3E6/Data/humanconnectome/Q3/151526/T1w/Diffusion.bedpostX/monitor: >>>>>>> >>>>>>> 10: [: -eq: unexpected operator >>>>>>> >>>>>>> __________________________________________________________________________________________________________________________________________________ >>>>>>> >>>>>>> >>>>>>> Preprocessing, processing and postprocessing did not start. I >>>>>>> have no >>>>>>> idea what to do. >>>>>>> >>>>>>> Best, >>>>>>> Rene >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>> >>>>>