You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I create this issue to share the problem encountered in ticket #11526 (https://elsa.onera.fr/issues/11526#change-379348).
It seems that the problem comes from Converter.Mpi methods. When you give a wrong file name in methods such as 'convertFile2PyTree' or 'convertFile2SkeletonTree', the calculation crashes but the job still runs.
The deadlock likely comes from the fact that processor 0 reads the file, then transmits the PyTree to other processors. With a wrong filename, processor 0 returns an error and the other processors wait for the PyTree, without returning errors.
You can find here a small case reproducing the problem, you can launch it using 'job_juno.sh' : /stck/ngruel/Public/11526
Thanks in advance,
Nicolas
The text was updated successfully, but these errors were encountered:
Hello everyone,
I create this issue to share the problem encountered in ticket #11526 (https://elsa.onera.fr/issues/11526#change-379348).
It seems that the problem comes from Converter.Mpi methods. When you give a wrong file name in methods such as 'convertFile2PyTree' or 'convertFile2SkeletonTree', the calculation crashes but the job still runs.
The deadlock likely comes from the fact that processor 0 reads the file, then transmits the PyTree to other processors. With a wrong filename, processor 0 returns an error and the other processors wait for the PyTree, without returning errors.
You can find here a small case reproducing the problem, you can launch it using 'job_juno.sh' : /stck/ngruel/Public/11526
Thanks in advance,
Nicolas
The text was updated successfully, but these errors were encountered: