When trying to reproduce the problem you mentioned, I found something similar but not exactly the same. I started with the single-mesh "roomfire" example problem and split it into two fire meshes and two evac meshes (and configured it for evac with a couple occupants). This worked fine when run in the normal (serial) way, it also worked fine in parallel with OpenMP, but when I tried to run in parallel with MPI I got this:
Starting FDS: C:\Program Files\PyroSim 2010\fds\fds5_mpi_win64.exe...
Process 3 of 3 is running on aurora.thunderheadeng.lan
Process 1 of 3 is running on aurora.thunderheadeng.lan
Process 0 of 3 is running on aurora.thunderheadeng.lan
Process 2 of 3 is running on aurora.thunderheadeng.lan
Mesh 1 is assigned to Process 0
Mesh 2 is assigned to Process 1
Mesh 3 is assigned to Process 2
Mesh 4 is assigned to Process 3
Fire Dynamics Simulator
Compilation Date : Fri, 29 Oct 2010
Version: 5.5.3; MPI Enabled; OpenMP Disabled
SVN Revision No. : 7031
Job TITLE : Room fire
Job ID string : roomfire
ERROR: EVACUATE_HUMANS, No mesh found 3
ERROR: EVACUATE_HUMANS, No mesh found 3
job aborted:
rank: node: exit code[: error message]
0: aurora.thunderheadeng.lan: 123
1: aurora.thunderheadeng.lan: 123
2: aurora.thunderheadeng.lan: 0: process 2 exited without calling finalize
3: aurora.thunderheadeng.lan: 123
I think there might be a problem with the way FDS+EVAC looks up the evacuation mesh data when all of the meshes have been assigned to different processes by MPI. However, that wasn't what you were seeing. I wasn't able to reproduce the PERS warning or your particular crash. Can you post your PSM or FDS input file?