You must have an active Palmetto account. To register for a Palmetto account, visit New Account Request page.
Step 1. Open OnDemand
- Open a browser and go to https://openod02.palmetto.clemson.edu/.
- Go to the
Step 2. Jupyter + Spark app
- Select the options for the app as described below, then click
Number of resource chunks (select):
CPU cores per chunk (ncpus):
Amount of memory per chunk (mem):
Number of GPUs per chunk (ngpus):
GPU Model (gpu_model):
1g - Ethernet older phases 1-6
Number of workers per node:
Only launch the driver on the master node.: Check this box.
- Once the job is ready and the infrastructure is launched, click
Connect to Jupyter.
Step 3. Getting Workshop Materials
- Run the following commands in the terminal
$ cp -R /zfs/citi/spark/myspark . $ cd myspark $ ls -l