Firefly and PC GAMESS-related discussion club



Learn how to ask questions correctly


FF internode fails

Jonas Baltrusaitis
jonas-baltrusaitis@uiowa.edu


I have no luck running ff between linux nodes. It runs fine on one node, no matter what I do I can't run it internode. This is my setup:

export P4_RSHCOMMAND=ssh in bash_profile
paswordless ssh
procgrp file with
local 3
mercury01 3
mercury02 3
mercury03 3

after all that I still get
Permission denied, please try again.

Permission denied, please try again.

Permission denied (publickey,password).

p0_15919:  p4_error: Child process exited while making connection to remote process on mercury01: 0

no idea how to solve that. I have FF folder on headnode and copy files to the scratch folder on headnode, maybe that's where the problem is?

cd $PBS_O_WORKDIR
set SCR=/scr/$USER/$PBS_JOBID

mkdir -p $SCR

set master=`hostname`
echo This job is running on host $master
echo under operating system `uname` at `date`
echo "Available scratch disk space (Kbyte units) at beginning of the job is"
df -k $SCR
if (-e $JOB) then
     cp $JOB $SCR/input
else
     echo "Input file $JOB does not exist. Please fix and resubmit."
endif
cp /home/jbaltrus/ff71g/firefly $SCR/
cp $PBS_O_WORKDIR/procgrp $SCR/
cp /home/jbaltrus/ff71g/fastdiag.ex $SCR/
cp /home/jbaltrus/ff71g/p4stuff.ex $SCR/

cp /home/jbaltrus/ff71g/pcgp2p.ex $SCR/
cd $SCR

./firefly -r -f -p -b -o $PBS_O_WORKDIR/output.out -p4pg procgrp  -ex /home/jbaltrus/ff71g

thanks

Jonas


[ Previous ] [ Next ] [ Index ]           Sat Jul 3 '10 4:17am
[ Reply ] [ Edit ] [ Delete ]           This message read 989 times