Search code examples
kubernetesdevopskubectl

"kubectl get all --all-namespaces" has different output against the same cluster

So when I run kubectl get all --all-namespaces on different machines, I get different output and I can't understand why.

First machine (output trimmed for brevity):

kube-system         po/kubernetes-dashboard-5569448c6d-w2bdb                                1/1       Running   0          16h
kube-system         po/service-mesh-traefik-5bb8d58bf6-gfdqd                                1/1       Running   0          2d
kube-system         po/tiller-deploy-78d74d4979-rh7nv                                       1/1       Running   0          23h

Second machine:

kube-system         kubernetes-dashboard-5569448c6d-w2bdb                                1/1       Running     0          16h
kube-system         service-mesh-traefik-5bb8d58bf6-gfdqd                                1/1       Running     0          2d
kube-system         tiller-deploy-78d74d4979-rh7nv                                       1/1       Running     0          23h

Ideally, I would like the output from the first machine, with the resource type prefixed to the ouput. Trying to comprehend the output of get all without it is extremely frustrating and due to the nature of the words in what I am searching for, it is even more frustrating trying to use Google to search for "get all".

What is different? The cluster is the same, so it should be returning the same data. The first machine is kubectl version 1.9.2, second machine is 1.10.0. The cluster is running 1.8.7.

Solution

  • The answer lies in the details you mentioned above. The difference in kubectl version displays in different output format between the two machines.