vllm - 💡(How to fix) Fix [Bug]: There should be flag/setting to collect environment [1 comments, 2 participants]

Fix Plan

To collect the environment using a Docker image without execing into the container, you can modify the Docker run command to run the collect_env.py script and output the results to a file that can be accessed outside the container.

Steps

• Run the Docker container with the collect_env.py script:

docker run -d --name vllm-container vllm/vllm-openai:v0.18.docker exec -d vllm-container wget https://raw.githubusercontent.com/vllm-project/vllm/main/vllm/collect_env.py -O /tmp/collect_env.py
docker exec -d vllm-container python /tmp/collect_env.py > /tmp/environment.txt

• Copy the environment file from the container to the host machine:

docker cp vllm-container:/tmp/environment.txt environment.txt

• Alternatively, you can use a Docker volume to mount a directory from the host machine to the container and output the environment file to that directory:

docker run -d --name vllm-container -v $(pwd):/tmp vllm/vllm-openai:v0.18.0
docker exec -d vllm-container wget https://raw.githubusercontent.com/vllm-project/vllm/main/vllm/collect_env.py -O /tmp/collect_env.py
docker exec -d vllm-container python /tmp/collect_env.py > /tmp/environment.txt

The environment file will be saved to the current directory on the host machine.

Verification

Verify that the environment file has been generated and contains the expected information:

cat environment.txt

This should display the collected environment information.