Failure in creating service when using "topologySpreadConstraints" in Deployment definition

Question

We are running an application inside a cluster we created in GKE. We have created required yamls (consisting of Service and Deployment definition). We recently have decided to use Pod Topology for that I have added following piece in my Deployment yaml file under spec section-

 spec:
  topologySpreadConstraints:
  - maxSkew: 1
    topologyKey: node
    whenUnsatisfiable: DoNotSchedule
    labelSelector:
      matchLabels:
        app: foo-app

This change is working as expected when I am running the service inside a minikube cluster while the same change is not working inside a GKE cluster. It throws an error-

Error: UPGRADE FAILED: error validating "": error validating data: ValidationError(Deployment.spec.template.spec): unknown field "topologySpreadConstraints" in io.k8s.api.core.v1.PodSpec

I searched a lot but could not find a satisfactory answer. Has anybody faced this problem? Please help me understand the problem and its resolution.

Thanks in Advance.

Got it. Client version was perfect (1.18.6) while server version was something which was causing problem. I upgraded that and it worked. Thanks :-) — Vinay Verma, Jul 29 '20 at 11:20

Neo Anderson · Accepted Answer · 2020-07-29T10:18:48.243

2

I assume you are running 1.17.7-gke.17 on your GKE cluster. Unfortunately this is the latest version you can upgrade to, through the rapid channel, at the time of this post.

topologySpreadConstraints is available in Kubernetes v1.18 FEATURE STATE: [beta]

edited Jul 29 '20 at 10:18

answered Jul 29 '20 at 10:13

Neo Anderson

5,957
2
12
29

1

Thanks Neo. I had to upgrade Server version (Client version was perfect `1.18.6`) and it worked. – Vinay Verma Jul 29 '20 at 11:21

Failure in creating service when using "topologySpreadConstraints" in Deployment definition

1 Answers1