kubectl-debug is an out-of-tree solution for troubleshooting running pods, which allows you to run a new container in running pods for debugging purpose. The new container will join the
ipc namespaces of the target container, so you can use arbitrary trouble-shooting tools without pre-install them in your production container image.
- quick start
- build from source
- default image and entrypoints
- future works
- implementation details
kubectl-debug is pretty simple, give it a try!
Install the debug agent DaemonSet in your cluster, which is responsible to run the "new container":
kubectl apply -f https://raw.githubusercontent.com/aylei/kubectl-debug/master/scripts/agent_daemonset.yml
Install the kubectl debug plugin:
# Linux curl -Lo kubectl-debug https://github.com/aylei/kubectl-debug/releases/download/0.0.1/kubectl-debug_0.0.1_linux-amd64 # MacOS curl -Lo kubectl-debug https://github.com/aylei/kubectl-debug/releases/download/0.0.1/kubectl-debug_0.0.1_macos-amd64 chmod +x ./kubectl-debug mv kubectdl-debug /usr/local/bin/
For windows user, download the latest binary from the release page and add it to your PATH.
Try it out!
kubectl debug POD_NAME # learn more with kubectl debug -h
Compatibility: I've tested
kubectl-debugwith kubectl v1.13.1 and kubernetes v1.9.1. I don't have an environment to test more versions but I suppose that
kubectl-debugis compatible with all versions of kubernetes and kubectl 1.12.0+. Please file an issue if you find
kubectl-debugdo not work.
Build from source
Clone this repo and:
# build plugin go build -o kubectl-debug ./cmd/plugin # install plugin mv kubectl-debug /usr/local/bin # build agent go build -o debug-agent ./cmd/agent # build agent image docker build . -t debug-agent
Default image and entrypoint
kubectl-debug use nicolaka/netshoot as the default image to run debug container, and use
bash as default entrypoint.
You can override the default image and entrypoint with cli flag, or even better, with config file
agent_port: 10027 image: nicolaka/netshoot:latest command: - '/bin/bash' - '-l'
kubectl-debug will always override the entrypoint of the container, which is by design to avoid users running an unwanted service by mistake(of course you can always do this explicitly).
kubectl-debug is supposed to be just a troubleshooting helper, and is going be replaced by the native
kubectl debug command when this proposal is implemented and merged in the future kubernetes release. But for now, there is still some works to do to improve
kubectl-debugruns privileged agent on every node, and client talks to the agent directly. A possible solution is introducing a central apiserver to do RBAC, which integrates to the kube apiserver using aggregation layer
kubectl-debugvendor the SPDY wrapper from
client-go. SPDY is deprecated now, websockets maybe a better choice
kubectl-debug consists of 2 components:
- the kubectl plugin: a cli client of
node agent, serves
- the node agent: responsible for manipulating the "debug container"; node agent will also act as a websockets relay for remote tty
When user run
kubectl debug target-pod -c <container-name> /bin/bash:
- The plugin get the pod info from apiserver and extract the
hostIP, if the target container is no existed or not currently running, an error raised.
- The plugin send a HTTP request to the specific node agent running on the
hostIP, which includes a protocol upgrade from HTTP to SPDY.
- The agent runs a container in the pod's namespaces (ipc, pid, network, etc) with the STDIN stay open (
- The agent checks if the target container is actively running, if not, write an error to client.
- The agent runs a
debug contaienrwill join the
usernamespace of the target container.
- The agent pipes the connection io to the
- Debug in the debug container.
- Jobs done, user close the SPDY connection.
- The node agent close the SPDY connection, then wait the
debug contaienrexit and do the cleanup.
Feel free to open issues and pull requests. Any feedback is highly appreciated!