Interactive Jobs on Discovery

Use the srun command to run jobs interactively on the Discovery HPC cluster. Visit the → Slurm Commands page for more details about the srun command.

Open OnDemand

  • One of the easiest ways to launch interactive jobs is through Open OnDemand. Please refer the documentation → Interactive Desktop to launch interactive desktop through which you can have an interactive command-line shell on compute node and run interactive jobs.

  • This will launch an interactive xfce session on the compute node and it’s highly recommended over the next method Interactive Shell on Compute Nodes Using Command Line which is discussed below.

Interactive Shell on Compute Nodes Using Command Line

To launch interactive shell on compute nodes using the command line, it’s recommended to use tmux and create a session first. If you accidentally quit your terminal or the connection to the discovery is lost, the changes or work you have done can be retrieved back again using tmux. Hence, first create a session using tmux and then run commands to launch interactive shell on compute nodes. To create/attach to a session using Tmux, please refer the following documentation → Tmux.

  • Use the srun command to run jobs interactively on the Discovery HPC cluster. Visit the → Slurm Commands page for more details about the srun command.

  • srun can also be used to run a login shell on the compute nodes which enables to run commands interactively. Consider the below example.

    srun -n 1 -c 2 --mem-per-cpu 5g  -p normal -t 01:00:00 --pty /bin/bash

  • Break down the parts of the command into a table format.

    Flags

    Alt

    Description

    -c

    --cpus-per-task

    The number of cpus to be used per task.

    --mem-per-cpu

    N/A

    The amount of RAM per CPU.

    -p

    --partition

    The partition to use.

    -t

    --time

    The execution wall time

    --pty

    /bin/bash

    The pseudoterminal (/bin/bash)

  • The command above simply means that you want to run a login shell (/bin/bash) on the compute nodes. When the command is executed, you’ll automatically get an interactive session on one of the compute nodes after which you can then run your commands interactively. You are also put into the working directory from which you ran the launched session.

    It’s crucial to specify the --pty flag for this to work as intended.

  • Now that you are logged into one of the compute nodes. Hence, whatever command you execute, for Example, srun hostname will be executed once (-n 1) using 2 cpus, based on the value of the srun flags declared above.

  • You can keep executing as many commands as you want. However, bear in mind that your interactive session on the compute node you are currently on, will be killed when the Walltime specified above elapses. Please see the example output below.

Example

  1. Log in to Discovery and execute the command below

    srun --ntasks 1 --cpus-per-task 2 --mem-per-cpu 10g -p normal -t 01:00:00 --pty /bin/bash

  2. You’d notice that the hostname changed from [username@discovery-l1 ~]$ to [username@discovery-g9 ~]$. This means you’re currently logged in to one of the compute nodes and now you can start executing jobs interactively. Next, execute the command below which prints the hostname of the node that you’re currently logged in to.

    echo "Hello from $(hostname)"
    *Output*
    Hello from discovery-g9.cluster.local

    From the above image, one can infer that the session switched from the login node to one of the compute nodes after the srun command gets executed. You can also see the output of the second command executed on the compute node Hello from discovery-g9.cluster.local.

    For more details on other useful srun flags please visit Slurm’s documentation on srun here https://slurm.schedmd.com/srun.html