Hi All, How does the hold job (or rerun) request work in a parallel job scenario (MPI or such programs started from inside of Torque)? I am interested in knowing the MOM side communications of this. Any help please? Thanks, Prakash