Following this tutorial ( https://portworx.com/run-ha-kafka-azure-kubernetes-service/) to setup a Kafka cluster on Azure Kubernetes Service, I am running in an issue deploying the Headless Zookeeper service.
Running this yml with kubectl to deploy the Zookeeper service, I get the error below.
apiVersion: v1
kind: Service
metadata:
name: zk-headless
nodeAffinity: labels:
app: zk-headless
spec:
ports:
- port: 2888
name: server
- port: 3888
name: leader-election
clusterIP: None
selector:
app: zk
error converting YAML to JSON: yaml: line 5: mapping values are not allowed in this contextMy question is, how do I assign the nodeAffinity value? Should I use this:
requiredDuringSchedulingIgnoredDuringExecution:
nodeSelectorTerms:In addition to Utku's answer - I would recommend using "kubectl explain service" command to see what fields can be added to a kubernetes object like service or deployment or Pod
Kubernetes Service objects do not have a nodeAffinity field, since Service resources do not run on nodes. In fact, they don't "run" at all - they are rather rules for the networking.
The example provided on the website has a copy-paste error there. It should have been:
apiVersion: v1
kind: Service
metadata:
name: zk-headless
labels:
app: zk-headless
spec:
ports:
- port: 2888
name: server
- port: 3888
name: leader-election
clusterIP: None
selector:
app: zk