April 2026

Patching a kubernets pod/container by mounting a script ontop of another one

It is possible to change binaries or scripts inside a pod during testing. The "kubectl cp" allows you to copy files into a running container if the permissions are configured such that you can do this.

The problem is however that such changes are temporary. If the container restarts then those chnages are gone. A kubernetes pod could be designed such that a container restart is part of normal error handling. So here is a way to easily patch the a container of a kubernets pod without building a complete new image for the pod.

The idea is that we create a config map containing the file you would like to change. This could e.g be a shell script saved inside a config map. It can also be a binary file but config maps have a size limit. From this config map we create a volume and this we mount the file from the volume on top of any existing script that we want to replace. It's a very simple process and to do this you just have to create the config map and edit the deployment of the container inside the pod.

Step 1) create the config map that contains the new script (or executable)

option 1) Use a text editor and creat a yaml file with this syntax:
```
apiVersion: v1
kind: ConfigMap
metadata:
  name: guido-patch-cm
data:
  modified-code.sh: |
    #!/bin/bash
    echo "this is the new code"
    date
```
Check that it created the config map:
kubectl get cm -A | grep guido-patch

option 2) create the config map from an existing script or binary file:
kubectl create configmap -n mynamespace guido-patch-cm --from-file=modified-code.sh

To run the above command you have to be in the directory where the script (modified-code.sh) is.
Step 2) edit the deployment of your pod

I suggest that you first save the existing deployment config file into a file and study it in detail. Complex pods could have big deployment configurations and it could be confusing if you try to edit them directly.
Find the existing deployment corresponding to your pod:
kubectl get deploy -A | grep mypod
Save it to a file:
kubectl get deploy -n mynamespace mypod -o yaml > mypod-deployment.yaml

To do the actual edit of the deployment with the vi editor you run the command:
kubectl edit deploy -n mynamespace mypod
After saving the changes and exiting the editor the pods will automatically restart and activate the new deployment configuration. You can then monitor the startup of the pod with:
kubectl get pods -A -o wide | grep mypod
```
apiVersion: apps/v1
kind: Deployment
metadata:
  name: mypod
....
spec:
  containers:
  - name: mypod-container
    ....more stuff here....
    volumeMounts:
    - name: script-volume
      mountPath: /opt/modified-code.sh  # Mount directly ontop of an existing file eclipsing the old one
      subPath: modified-code.sh         # Pick specific file from configmap
  volumes:
  - name: script-volume
    configMap:
      name: guido-patch-cm
      defaultMode: 0755          # Make the script executable
```

This metod allows you to eaily patch and change a pod while you are testing it and while it is running. It's much faster than installing a new build. Pemmican is the original survival food of the north. You can keep it for years without refrigeration but it is nutritionally complete. Unlike dried cookies or modern energy bars it does not spike sugar and deprive you of nutrients.