Add a svg demo file showing the DRA use: install, configure and use

[yuanchen8911]

This PR adds a svg file to README showing a complete example with all the steps outlined in the README on how to install, configure and use DRA to run a pod on a Linux workstation, including all the commands and output.

Hopefully, It can assist users in comparing and identifying issues as they play around.

https://github.com/yuanchen8911/k8s-dra-driver/blob/demo-video/demo/specs/quickstart/basic-demo.svg

Not sure where to place the svg file. Maybe we should create a separate folder for it (and future demo videos if any).

[yuanchen8911]

/cc @klueska @ArangoGutierrez

[yuanchen8911]

/cc @klueska
/cc @ArangoGutierrez

[klueska]

In general this change looks good now. My only worry is keeping this SVG in sync as we make changes to the README. Is there someway to add the code to generate it in the repo instead of just importing the final image?

[yuanchen8911]

Yes, but I assume we won't change the basic flow and example very frequently.
The svg file was generated from the following description script file using a tool I created. recdemo.sh <input: demo script> <output: svg file>.

We can add the script and demo files to a new folder and update/build svg files (e.g., in Makefile) when things change. That's how KWOK manages its examples and demos.

How about doing it in future PRs?

# Demonstrate the basic use of DRA in a kind cluster

# Prerequisites: follow the instructions in the README
# 1. Install kind
# 2. Install and configure the NVIDIA container toolkit

# Show the current set of GPUs on the machine
nvidia-smi -L

# Create a kind cluster to run the demo:
./demo/clusters/kind/create-cluster.sh

# Build the image for the example resource driver and the images for the kind cluster:
./demo/clusters/kind/build-dra-driver.sh

# Install the driver in the kind cluster
./demo/clusters/kind/install-dra-driver.sh

# Show two pods running in the nvidia-dra-driver namespace:
kubectl get pods -n nvidia-dra-driver

# Run the examples in the demo/specs/quickstart folder. The README in that directory shows the full script:
# For example, you can run the second test.
kubectl apply --filename=demo/specs/quickstart/gpu-test2.yaml
sleep 2

# Get the pod information
kubectl get pod -n gpu-test2

# Get the GPU resource information of the pod.
"basic.demo" 34L, 1109B                                                                                                      1,1           Top
# 2. Install and configure the NVIDIA container toolkit

# Show the current set of GPUs on the machine
nvidia-smi -L

# Create a kind cluster to run the demo:
./demo/clusters/kind/create-cluster.sh

# Build the image for the example resource driver and the images for the kind cluster:
./demo/clusters/kind/build-dra-driver.sh

# Install the driver in the kind cluster
./demo/clusters/kind/install-dra-driver.sh

# Show two pods running in the nvidia-dra-driver namespace:
kubectl get pods -n nvidia-dra-driver

# Run the examples in the demo/specs/quickstart folder. The README in that directory shows the full script:
# For example, you can run the second test.
kubectl apply --filename=demo/specs/quickstart/gpu-test2.yaml
sleep 2

# Get the pod information
kubectl get pod -n gpu-test2

# Get the GPU resource information of the pod.
kubectl logs -n gpu-test2 pod --all-containers

# Delete the kind cluster and clean up the environment
 ./demo/clusters/kind/delete-cluster.sh

[klueska]

OK. I'm fine merging it as-is for now.

[yuanchen8911]

@klueska I hope the current version has addressed all your comment. Can you take another look? Thanks!