Interactive Pods & Long-Idle Guidelines

Interactive pods enable live development, notebook prototyping, and terminal access to GPU hardware.

Launching a Development Pod

kubectl run dev-session \
  --namespace sci-myproject \
  --image=pytorch/pytorch:2.3.0-cuda12.1-cudnn8-runtime \
  --limits='nvidia.com/gpu=1,cpu=8,memory=32Gi' \
  --restart=Never \
  -it -- /bin/bash

Best Practices for Interactive Work

Clean up when done: Do not leave idle GPU pods running overnight without computations. When finished, exit and delete the pod:
Terminal window
```
kubectl delete pod dev-session -n sci-myproject
```
Mount Persistent Storage: Always mount a CephFS PVC so your notebooks, code edits, and conda environments persist across pod restarts.
Port Forwarding for Web UIs: If your pod runs TensorBoard, Streamlit, or JupyterLab on port 8888:
Terminal window
```
kubectl port-forward dev-session 8888:8888 -n sci-myproject
```
Open http://localhost:8888 on your local laptop to access the interactive web interface securely via encrypted SSH/tunnel.