[omniORB] omni_thread_fatal while using queuing engine

[Alexander K]

Hey guys,

I am using omniORB / CORBA via Salome CAD software. I wrote a script which
is called with following command "runAppli -t myscript.py". runAppli is the
executable of Salome. My script generates a geometry according to given
design parameters and meshes it.

If I call my script with "runAppli -t myscript.py" on my workstation
everything runs smooth. BUT if I do the same and submit this into a queing
system (SGE) something like "qsub runAppli -t myscript.py" salome crashes
with the message:

terminate called after throwing an instance of 'omni_thread_fatal'



/home/username/salome/appli_V7_2_0/runAppli: line 50: 19895 Aborted
(core dumped) ${KERNEL_ROOT_DIR}/bin/salome/envSalome.py python
${KERNEL_ROOT_DIR}/bin/salome/runSalome.py $*

So 'omni_thread_fatal' gives me the clue that it has to do something with
the omniORB stuff. I tried a lot with the "qsub" command forwarding the
environment variables (-V) but that's somehow not the problem I think.

The nodes in the queuing grid have the same home directory mounted and I
prepared a python 2.7 and a salome 7.2.0 version in the home directory. So
it really is available on each node in the grid. The environment variables
seem good to me (comparing output of "env" & "qsub env")

Has anybody else tried that before and experienced the same issues? And is
there a solution for that? Hope somebody can help me out here. What could
cause such a problem?


Thanks in advance...
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.omniorb-support.com/pipermail/omniorb-list/attachments/20140102/29f0dc97/attachment.html>