MPI crashes with the error message: (it_plfd->revents & POLLERR) == 0

[seongyun_k_]

Does any one know what the following error message mean? where should I check?

[20] rank = 0, revents = 24, state = 8

===================================================================================
=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
=   PID 11769 RUNNING AT XXXX
=   EXIT CODE: 9
=   CLEANING UP REMAINING PROCESSES
=   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================
[20] Assertion failed in file ../../src/mpid/ch3/channels/nemesis/netmod/tcp/socksm.c at line 2969: (it_plfd->revents & POLLERR) == 0
[20] internal ABORT - process 20
[20]
[20]
[20]
[20]

[James_T_Intel]

How are you launching your job?  Code 9 usually indicates the process was killed externally (CTRL-C, SIGKILL, etc.).  Is this repeatable?  Can you run with I_MPI_HYDRA_DEBUG=1 and post the output (as a file please)?