When I use "kubectl patch" to delete a specific object in an array, I need to first check the index of that specific object, and then do the kubectl patch with the json/yaml path.
For example, I have an resource contains multiple port object:
spec:
clusterIP: 100.12.143.76
externalTrafficPolicy: Cluster
ports:
- name: object1
nodePort: 30681
port: 1234
protocol: TCP
targetPort: 1234
- name: object2
nodePort: 31805
port: 9876
protocol: TCP
targetPort: 9876
If I need to remove the "object1" in the ports, here is the command I am using now (kong-proxy is the name of this service resource):
kubectl patch service kong-proxy --type=json -p='[{"op": "remove", "path": "/spec/ports/0"}]' -n kong
By doing the command above, it deletes the object at index "0" which with the name "object1".
But this requires the person to know the correct index number in advance, and if there are too many objects in an array, it will be hard to manage and calculate the index.
Therefore I wonder without writing a separate bash script, is there a way that I can delete the object with an identifier in it.
Such as:
kubectl patch service kong-proxy --type=json -p='[{"op": "remove", "path": "/spec/ports/{name:object1}"}]' -n kong
I'm afraid it's not possible. The kubectl patch --type=json
command uses JSON Patch under the hood, which in turn uses JSON Pointer. JSON Pointer standard is pretty simple and does not provide filtering/matching functionality.
What you could do as a work around is using jq to find the array index, before calling kubectl patch
:
INDEX=$(kubectl get svc kong-proxy -n kong -o json | jq '.spec.ports | map(.name == "object1") | index(true)')
kubectl patch svc kong-proxy --type=json -p="[{'op': 'remove', 'path': '/spec/ports/$INDEX'}]" -n kong