To forward a port using srun in a Slurm environment, you typically want to set up an SSH tunnel to a compute node where your job is running. Here’s how you can do it:

---

1. Request an Interactive Job with Port Forwarding

When you submit an interactive job with srun, you can request port forwarding using the --tcpip or --tunnel options, but Slurm itself does not directly support port forwarding. Instead, you usually combine srun with SSH tunneling.

Basic Command:

bash
Copy
srun --pty --nodes=1 --ntasks-per-node=1 --cpus-per-task=1 --mem=4G --time=01:00:00 --x11=first bash

• This gives you an interactive shell on a compute node.

---

2. Set Up SSH Tunneling

Once your job is running, you can set up an SSH tunnel from your local machine to the compute node.

Find the Compute Node:

When your srun job starts, it will print the hostname of the compute node, e.g., node123.

Create the SSH Tunnel:

On your local machine, run:

bash
Copy
ssh -L <local_port>:localhost:<remote_port> <your_username>@<login_node> -J <your_username>@<compute_node>

• Replace:
  • <local_port>: The port on your local machine (e.g., 8888).
  • <remote_port>: The port on the compute node (e.g., 8888).
  • <your_username>: Your username on the cluster.
  • <login_node>: The login node of the cluster.
  • <compute_node>: The compute node where your job is running (e.g., node123).

Example:

bash
Copy
ssh -L 8888:localhost:8888 benjamin@login.cluster.edu -J benjamin@node123

• This forwards localhost:8888 on your machine to node123:8888.

---

3. Use the Forwarded Port

Now, any service running on <remote_port> (e.g., a Jupyter notebook) on the compute node will be accessible on your local machine at localhost:<local_port>.

---

Alternative: Use srun with --tunnel (if supported)

Some Slurm setups allow you to specify --tunnel directly with srun:

bash
Copy
srun --tunnel=8888:8888 --pty bash

• This is less common and depends on your cluster’s configuration.

---

Notes:

• Ensure the port you want to forward is not blocked by the cluster’s firewall.
• If you’re running a service (e.g., Jupyter), make sure it’s bound to 0.0.0.0 or the node’s hostname, not just localhost.

---