Connecting to the ParaView data server on lobo will require two cygwin terminals and the ParaView 3.4 client. To make life easier when tunneling between the workstation and the server we will use cygwin, Start->All Programs->Cygwin. However if you wan to use putty see the section on how to configure , here. In the first cygwin terminal, we will login to the lobo front end in two hops, launch the interactive job, load the environment modules and, finally start the ParaView data server. In the second cygwin terminal we will login into the compute node assigned us by the batch system in three hops, and create the ssh tunnel. Only then will we be able to connect the ParaView client running on our workstation to the data server running on the compute node.

Open the first cygwin terminal and enter the following sequence of commands:

1$ ssh username@
1$ ssh lo1-fe
1$ llogin -n 32


At this point you should see a message indicating that the interactive job has started and be at a login shell on one of the compute nodes, lobXXX, where XXX will be a node number chosen by the batch system. Before starting the ParaView data server we will open the second cygwin terminal and establish the tunnel to the compute node which for the sake of this example we will call lob103. Open the second cygwin terminal and enter the following commands:

2$ ssh -L 11111:localhost:44444 homakar@
2$ ssh -L 44444:localhost:44444 lo1-fe
2$ ssh -L 44444:localhost:44444 lob103

This creates the ssh tunnel. Now switch back to the first cygwin terminal, to the waiting interactive batch job and enter the following commands to start the ParaView data server:

1$ setenv MODULEPATH ${MODULEPATH}:/usr/projects/views/modulefiles
1$ module load paraview/3.4.0-lobo-Mesa7.3
1$ module load openmpi-gcc/1.2.8
1$ mpirun -n 32 pvserver --use-offscreen-rendering

Now on your workstation Start->All Programs->ParaView->ParaView-3.4, and connect using port 11111. If you haven't configured a connection on your workstation to port 11111 yet refer to the section on how to do so, here The VPIC plugin should have loaded automatically on both the client and data server. To verify one can open Tools->manage Plugins/Extensions dialog and look for the VPIC entries. At this point you should be able to browse the server's file system and open datasets.