Issue Description
can not be start twice
Type: bug report
Describe what happened
when start the gpu-docker-api service ,but there is already runing one
Describe what you expected to happen
after restart the service the available gpu must sync with the really env, and update the etcd date in the startup step
How to reproduce it (as minimally and precisely as possible)
- start api
- using api create one container request some gpus
- start the api again
- using api to delete the container
Tell us your environment
ubuntu 22.04
Anything else we need to know?
Issue Description
can not be start twice
Type: bug report
Describe what happened
when start the gpu-docker-api service ,but there is already runing one
Describe what you expected to happen
after restart the service the available gpu must sync with the really env, and update the etcd date in the startup step
How to reproduce it (as minimally and precisely as possible)
Tell us your environment
ubuntu 22.04
Anything else we need to know?