Setup

Log into the HPC cluster’s On Demand interface

• Open a Chrome browser and navigate to the OnDemand Interface
• Log in with your Tufts Credentials
• On the top menu bar choose Clusters -> Tufts HPC Shell Access

• You’ll see a welcome message and a bash prompt, for example for user tutln01:

[tutln01@login001 ~]$

• This indicates you are logged in to the login node of the cluster. Type clear to clear the screen

Storage Space

• Check how much available storage you have in your home directory by typing showquota:

Home Directory Quota
Disk quotas for user tutln01 (uid 31394):
     Filesystem  blocks   quota   limit   grace   files   quota   limit   grace
hpcstore03:/hpc_home/home
                  1222M   5120M   5120M            2161   4295m   4295m        


Listing quotas for all groups you are a member of
Group: facstaff	Usage: 16819478240KB	Quota: 214748364800KB	Percent Used: 7.00%

• What does this mean?

  • blocks: amount of storage you are using. Here it is 1222M
  • quota: amout of available storage. Here it is 5120M

NOTE: If you do not have 500M available, you may have space in a project directory for your lab. These are located in /cluster/tufts with names like /cluster/tufts/labname/username/. If you don’t know whether you have project space, please email tts-research@tufts.edu.

Download the data

• As of now we are on a login node. We CANNOT run our scripts on the login node. For that we need a compute node, and we can request one like so:

srun --pty -t 3:00:00 --mem 16G -N 1 --cpus 4 bash

Where:

• srun requests the pseudo terminal
• --pty bash requests a bash terminal
• -t 3:00:00 requests 3 hours on that compute node
• --mem 16G requests 16G of memory
• -N 1 requests 1 compute node
• --cpus 4 requests 4 cpus on that 1 compute node

NOTE: If wait times are very long, you can try a different partitions by adding, e.g. -p preempt or -p interactive before bash. If you go through this workshop in multiple steps, you will have to rerun this step each time you log in.

• Change to your home directory

  cd

  Or, if you are using a project directory:

  cd /cluster/tufts/labname/username/

• Copy the course directory and all files in the directory (-R is for recursive):

  cp -R /cluster/tufts/bio/tools/training/microbiome16S/ .

Working with an Interactive App

Today we will be working with R. We can access R using the RStudio interactive app OnDemand. To do this we will:

• Go to Interactive Apps -> RStudio Pax

• Type in the number of hours you’ll use the app for, how memory you’ll need, what version of R you’ll need, the reseveration, and any supporting modules. Here enter the following:
  • hours: 4
  • memory: 16G
  • R version: 4.0.0
  • reservation: Default
  • supporting modules: don’t enter anything here
• Click Launch and an RStudio window will pop up

---

Next

Previous