[gmx-users] Reg Gromacs in cluster problem

[vidhya sankar]

Dear Justin,
                     Thank you for your Previous reply  I am Running Gromacs on Cluster  with 16 Processor
I am using the following script to submit Job

#!/bin/bash

#PBS -l nodes=compute-0-5:ppn=16
#PBS -l walltime=900:10:5
hostname
date


cd $PBS_O_WORKDIR
echo "files copied from" $PBS_O_WORKDIR
echo "to computing directory" $TMPDIR
cd $TMPDIR


cp $PBS_O_WORKDIR/newplumed04RestartHILLS.dat  $TMPDIR/

mpirun=/opt/openmpi/bin/mpirun

LD_LIBRARY_PATH=/share/apps/gromacsplu/lib

source="/share/apps/gromacsplu//bin"
MDRUN="/share/apps/gromacsplu/bin/mdrun_mpi_d"
$mpirun -np 16 $MDRUN -s CNTPEPRSOLIONSfullplumed.tpr -npme 0 -cpi CNTPEPRSOLIONSfullplumed_prev.cpt -append  -nt 1  -plumed newplumed04RestartHILLS.dat -v -deffnm CNTPEPRSOLIONSfullplumed &>/dev/null

cp --force $TMPDIR/* $PBS_O_WORKDIR/out14/

rm -rf $TMPDIR
date


But after 19 hours Running My Node Become down  and shows as  follows

state = down,job-exclusive
     np = 16
     ntype = cluster


When i Try to Access My node using the following Command 

ssh compute-0-5   then it shows the follwong error

ssh: connect to host compute-0-5 port 22: No route to host

I would be very garteful if you Give any suggestion
 
Thanks in Advance