Kubernetes开发新手：如何向Kube-proxy推送代码变更？

Hey there! As someone who’s tinkered with customizing kube-proxy on ARM clusters before, I can help you streamline this workflow so you don’t have to manually swap binaries on each Raspberry Pi every time you make a change. Let’s break this down into practical, actionable steps:

First, quick context: kube-proxy is almost always deployed as a DaemonSet in the kube-system namespace, meaning it runs on every node in your cluster. Manually replacing binaries works for one-off tests, but using Kubernetes’ native tools will make your changes reproducible and easier to roll back.

1. Build a Custom ARM-Compatible Kube-proxy Image #

Since Raspberry Pis use ARM architecture, you’ll need to cross-compile your modified kube-proxy binary and package it into a Docker image that your RPI nodes can run. Here’s how to do it with Kubernetes’ official build tools:

• From your local dev machine (where you already cross-compile), run this command to build the ARM64 binary and image:

  make WHAT=cmd/kube-proxy ARCH=arm64 REGISTRY=your-docker-username TAG=v1.28.0-custom
  

  Replace your-docker-username with your Docker Hub handle (or a local private registry if you don’t want public pushes) and v1.28.0-custom with a tag that matches your cluster’s Kubernetes version plus a custom suffix.

• If your RPIs don’t have internet access, skip pushing to a registry and instead save/load the image locally:

  docker save your-docker-username/kube-proxy:v1.28.0-custom -o kube-proxy-custom.tar
  # Copy to each node and load the image
  for node in rpi-node-1 rpi-node-2 rpi-node-3 rpi-node-4; do
    scp kube-proxy-custom.tar pi@$node:/tmp/
    ssh pi@$node "docker load -i /tmp/kube-proxy-custom.tar"
  done
  

2. Update the Kube-proxy DaemonSet to Use Your Custom Image #

Tell Kubernetes to replace the default kube-proxy with your modified version:

• Edit the DaemonSet directly (fastest for quick changes):

  kubectl edit daemonset kube-proxy -n kube-system
  

  Find the containers[0].image field and swap the default image with your custom one (e.g., your-docker-username/kube-proxy:v1.28.0-custom). Save and exit—Kubernetes will automatically roll out new Pods to all nodes.

• For version-controlled changes, export the current config, modify it, then apply:

  kubectl get daemonset kube-proxy -n kube-system -o yaml > kube-proxy-custom.yaml
  

  Update the spec.template.spec.containers[0].image line in the YAML, then apply:

  kubectl apply -f kube-proxy-custom.yaml -n kube-system
  

3. Verify Your Changes Are Working #

After deployment, confirm everything’s running as expected:

• Check Pod status across all nodes:

  kubectl get pods -n kube-system -l k8s-app=kube-proxy
  

• View logs to ensure your custom logic is executing:

  kubectl logs -n kube-system -l k8s-app=kube-proxy
  

• Exec into a Pod to verify the binary version or test functionality:

  kubectl exec -n kube-system $(kubectl get pods -n kube-system -l k8s-app=kube-proxy -o name | head -1) -- /usr/local/bin/kube-proxy --version
  

For Quick, Temporary Tests: Batch Binary Replacement #

If you still want to stick with binary swaps for fast iterations (e.g., testing tiny code changes without building an image), use a script to automate the process across all 4 RPIs:

Create a shell script named update-kube-proxy.sh:

#!/bin/bash
# Path to your cross-compiled kube-proxy binary
LOCAL_BINARY="./build/bin/kube-proxy"
# List of your RPI nodes (update with your actual hostnames/IPs)
NODES=("rpi-node-1" "rpi-node-2" "rpi-node-3" "rpi-node-4")

for node in "${NODES[@]}"; do
  echo "Updating kube-proxy on $node..."
  scp "$LOCAL_BINARY" pi@$node:/tmp/
  ssh pi@$node "sudo systemctl stop kube-proxy && sudo cp /tmp/kube-proxy /usr/local/bin/kube-proxy && sudo systemctl start kube-proxy"
done

echo "All nodes updated!"

Make it executable with chmod +x update-kube-proxy.sh and run it whenever you need to push changes.

Quick Tips to Avoid Headaches #

• Test changes on one node first (use node selectors in the DaemonSet) before rolling out to all 4 RPIs.
• Ensure your custom kube-proxy version matches your cluster’s Kubernetes version to avoid compatibility bugs.
• If you’re modifying kube-proxy’s configuration (not just code), don’t forget to update the kube-proxy-config ConfigMap in kube-system if needed.

内容的提问来源于stack exchange，提问作者elia