gyral parcellation - Run in parallel

[joost]

Hi,

stamp: BV version 4.2.1, Ubuntu 64-bits.
i'm trying to use the Gyral Parcellation process (Cortical Surface->old) using SOMA (Run in parallel). this is the output given for 1 subject under the 'standard output' tab:

Loading toolbox RII

Loading toolbox brainrat

Loading toolbox connectomist

Loading toolbox cortical_surface

Loading toolbox data management

Loading toolbox datamind

Loading toolbox editors

Loading toolbox freesurfer

Loading toolbox morphometry

Loading toolbox structural_analysis

Loading toolbox sulci

Loading toolbox t1mri

Loading toolbox tests

Loading toolbox tools

Loading toolbox viewers

Loading toolbox My processes

ERROR No database contain a DiskItem with uuid

172c4168-a9d3-dc41-464c-1226ad07c19c



--------------------------------------------------------------------------------

DatabaseError

neuro.py (259) in main:



exec f in localsStartup, localsStartup



<string> (1) in <module>:



?



neuroProcesses.py (2170) in runProcess:



_process = getProcessInstance( _process )



neuroProcesses.py (3418) in getProcessInstance:



result = result()



neuroProcesses.py (938) in __init__:



Parameterized.__init__( self, self.signature.shallowCopy())



neuroProcesses.py (539) in __init__:



self.initialization()



2DWhiteMatterParcellation.py (123) in initialization:



?



sqlFSODatabase.py (1514) in getDiskItemFromUuid:



?

i have updated my database but the error persists. other processing (e.g. Morphologist) parallelizes fine. what am i doing wrong?

merci,
-joost

[Dominique Geffroy]

Hello,

I'm affraid this process cannot be run in parallel for the moment. Indeed, it seems to try accessing the Brainvisa database whereas the databases features are disable in parallel run mode because because our database system cannot manage concurrent access.
Maybe it is possible to change this process to avoid accessing the database, we'll look at that for the next version, but for the moment, you should run it in normal mode.

Dominique

[riviere]

Actually the database access was only used during parameters initialization, which is useless in distributed execution mode since all parameters are completely specified. I have fixed it for the next version: database should not be mandatory.
Denis