[mvapich-discuss] MPIDI_CH3I_SMP_Init(1852): write: Success
吕秋义
lvqiuyi at 126.com
Fri Jan 6 04:27:07 EST 2017
Hello,
I want to run Gromacs in Docker using mvapich2-2.1. I have installed infiniband in my host and I have used the same software running in the host well. But when I run Gromacs in Docker I got an error:
connect [mt_checkin]: Connection refused
connect [mt_checkin]: Connection refused
connect [mt_checkin]: Connection refused
[infiniband-mynode2:mpirun_rsh][child_handler] Error in init phase, aborting! (1/12 mpispawn connections)
connect [mt_checkin]: Connection refused
[infiniband-mynode2:mpirun_rsh][child_handler] Error in init phase, aborting! (1/12 mpispawn connections)
connect [mt_checkin]: Connection refused
connect [mt_checkin]: Connection refused
The command I use is "mpirun_rsh -np 192 -hostfile /home/lqy/new/host /home/lqy/gromacs-4.5.3/bin/mdrun_mpi -s /home/lqy/new/lmd_10.tpr -deffnm lmd_10"
When I use the command "mpirun -np 192 -hostfile /home/lqy/new/host /home/lqy/gromacs-4.5.3/bin/mdrun_mpi -s /home/lqy/new/lmd_10.tpr -deffnm lmd_10". I got the following error:
[cli_0]: aborting job:
Fatal error in MPI_Init:
Other MPI error, error stack:
MPIR_Init_thread(514)....:
MPID_Init(359)...........: channel initialization failed
MPIDI_CH3_Init(446)......:
MPIDI_CH3I_SMP_Init(1852): write: Success
[infiniband-mynode2:mpi_rank_144][error_sighandler] Caught error: Segmentation fault (signal 11)
[infiniband-mynode2:mpi_rank_84][error_sighandler] Caught error: Segmentation fault (signal 11)
[infiniband-mynode2:mpi_rank_108][error_sighandler] Caught error: Segmentation fault (signal 11)
[infiniband-mynode2:mpi_rank_36][error_sighandler] Caught error: Segmentation fault (signal 11)
[infiniband-mynode2:mpi_rank_60][error_sighandler] Caught error: Segmentation fault (signal 11)
[infiniband-mynode2:mpi_rank_96][error_sighandler] Caught error: Segmentation fault (signal 11)
[infiniband-mynode2:mpi_rank_48][error_sighandler] Caught error: Segmentation fault (signal 11)
[infiniband-mynode2:mpi_rank_156][error_sighandler] Caught error: Segmentation fault (signal 11)
[infiniband-mynode2:mpi_rank_72][error_sighandler] Caught error: Segmentation fault (signal 11)
I use ipoib and I use pipework to pass the IP of the IB to the container.
I can't find any information about this error. Any help would be appreciate!!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cse.ohio-state.edu/pipermail/mvapich-discuss/attachments/20170106/8794b658/attachment.html>
More information about the mvapich-discuss
mailing list