Hello

I have a question.

I have a condor pool.

Can I do run the seti@home in my condor claster system?

Thank you

---

I also have access to a condor cluster, as well as a 100 node HPCC cluster running Platform OCS/PBS.

Is there a platform ocs role or instructions for pbs?

Regards,

Brett

---

I have readen in a other thread, that peoples has successfull distributed single BOINC instances via HPC mechanisms to the nodes and started a dedicated BOINC instance on each node.

---

Hi again

I'm woking the condor problem.

But.

I have an another idea.

So..

Step 1

I created the NFS home in all PC in the claster.

NFS server:
/etc/export:
/home 10.0.1.0/255.255.255.0(rw,no_subtree_check,sync,no_root_squash)

Workstation:
/etc/ftsab:
10.0.1.1:/home /home nfs defaults,nolock 0 2

Mount, home or reboot workstation machine after add your user on /etc/passwd autghenticatin, or add LDAP+Kerberos, etc...

Warning: Mind the permissions!

Step 2

I created the ssh key..

ssh-keygen -t dsa ( The passphrase is emply! )
cp ~/.ssh/id_dsa.pub ~/.ssh/authorized_keys2

Step 3

mkdir ~/projects/nodes/HOSTNAME1 ( Sample: mkdir ~/projects/nodes/node75 )
mkdir ~/projects/nodes/HOSTNAME2
mkdir ~/projects/commands
mkdir ~/projects/tools

Step 4
( My claster members is 14 ) If you have node host make this the all host

cd ~/projects/nodes/HOSTNAME1
wget http://boincdl.ssl.berkeley.edu/dl/boinc_5.10.21_i686-pc-linux-gnu.sh
sh boinc_5.10.21_i686-pc-linux-gnu.sh
cp -v ./boinc_5.10.21_i686-pc-linux-gnu.sh ~/projects/tools/

All hosts!

Step 5

cd ~/projects/nodes/HOSTNAME1/BOINC
./boinc --attach_project http://setiathome.berkeley.edu/ xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ( xxxxx*X = Account key )

All directory!

Step 6

Wait the attact project proccess

Step 7

CTRL+C

Step 8

cd ~/projects/commands

cat > boincctl << "EOF"
#!/bin/bash

# created: Csiga | http://setiathome.berkeley.edu/show_user.php?userid=53878
case $1 in
start)
echo "Starting BOINC nodes in claster..."
ssh HOSTNAME1 ~/projects/nodes/HOSTNAME/BOINC/start.sh
ssh HOSTNAME2 ~/projects/nodes/HOSTNAME/BOINC/start.sh
;;

stop)
echo "Stopping BOINC node in claster..."
ssh HOSTNAME1 killall boinc
ssh HOSTNAME2 killall boinc
;;
status)
echo "-------------------"
echo "Process HOSTNAME1:"
ssh HOSTNAME1 ~/projects/commands/run
echo "-------------------"
echo "Process HOSTNAME2:"
ssh HOSTNAME2 ~/projects/commands/run
echo "-------------------"
;;
*)
echo "Usage: $0 {start|stop|status}"
exit 1
;;
esac

EOF

Step 9

chmod +x boincctl


Step 10

Create the start.sh files in your hosts directoryes:

cd cd ~/projects/nodes/HOSTNAME1/BOINC
cat > start.hu << "EOF"
cd "~/projects/nodes/HOSTNAME1/BOINC" && exec ./boinc --daemon $@

EOF

chmod +x start.sh

All host directory!

Step 11

Make the "run" file:

cat > run << "EOF"
#!/bin/bash

PS=/bin/ps
GREP=/bin/grep
SH=/bin/sh
ECHO=/bin/echo

$PS aux -www | $GREP setiathome | $GREP -q -v grep
if [ $? = 0 ]; then
echo "Running"
else
echo "Not running"
fi

EOF
chmod + run

Step 12

Lets go!

Login the NFS server, or any node and:

Starting:
cd ~/projects/commands
./boincctl start

Stopping:

cd ~/projects/commands
./boincctl stop

Status:
cd ~/projects/commands
./boincctl status

Good luck!

---