bitfusion-with-kubernetes-integration
bitfusion-with-kubernetes-integration copied to clipboard
support container entrypoint
Please consider adding support for configuring a kubernetes container command with bitfusion based on the container image entrypoint. Currently only pods that specify a container command do work.
I would also like to see support for a pod without a command: [] in the spec supported
The use case is deploying a jupyterhub server on Kubernetes. Jupyter hub deploys jupyter notebooks at scale and GPU support in these would be quite useful to many Jupyterhub run a python app that starts pods that look like this (note only args is present, no command is present in the pod spec):
"containers": [
{
"args": [
"jupyterhub-singleuser",
"--ip=0.0.0.0",
"--port=8888"
],
hello, expert. I did a poc follow this guide, but I could't create benchmark pod at step4. I got a message from the describe pod. pls refer to the message
Warning Failed 30m (x5 over 32m) kubelet, 10.168.47.36 Error: failed to start container "bf-pkgs": Error response from daemon: OCI runtime create failed: container_linux.go:380: starting container process caused: process_linux.go:545: container init caused: Running hook #0:: error running hook: exit status 1, stdout: , stderr: nvidia-container-cli: initialization error: driver error: failed to process request: unknown Warning BackOff 2m16s (x133 over 31m) kubelet, 10.168.47.36 Back-off restarting failed container pls , what docker version and libary ?