Today, when I deployed fcp on a machine without a GPU, I deployed it according to the documentation, but in the end the ubi task pod could not start normally. Later I found that the environment variables in fil-c2.env needed to be changed.
Since the deployed machine does not have a GPU, fil-c2.env needs to add BELLMAN_NO_GPU=1 instead of using the RUST_GPU_TOOLS_CUSTOM_GPU
Thanks for your feedback. We have fixed the issue in the releases branch, When you receive a CPU task, the BELLMAN_NO_GPU=1 will be set in the fil-c2 task.
Today, when I deployed fcp on a machine without a GPU, I deployed it according to the documentation, but in the end the ubi task pod could not start normally. Later I found that the environment variables in fil-c2.env needed to be changed.
Since the deployed machine does not have a GPU, fil-c2.env needs to add BELLMAN_NO_GPU=1 instead of using the RUST_GPU_TOOLS_CUSTOM_GPU