HA in azure kubernetes using Podantiaffinity

Question

Hello,

To achieve HA for my application (3 replicas) do I need to use the podantiaffinity as preferredDuringSchedulingIgnoredDuringExecution or requiredDuringSchedulingIgnoredDuringExecution? Also if I go with topologykey = zone can I do a new deployment? we are using availability zone1, zone2 and zone3

preferredDuringSchedulingIgnoredDuringExecution:
            - labelSelector:
                matchExpressions:
                  - key: app.kubernetes.io/name
                    operator: In
                    values:
                      - abcd
              topologyKey: topology.kubernetes.io/zone

Answer 1

@arsh ,

That is correct! There is one more feature called "topology spread constraints" - that should also help out in evenly distributing your pods managed by deployment object on the nodes!

Take a look at the below article how to implement that:

https://www.danielstechblog.io/distribute-your-application-across-different-availability-zones-in-aks-using-pod-topology-spread-constraints/

Kubernetes official documentation about that feature:

https://kubernetes.io/docs/concepts/scheduling-eviction/topology-spread-constraints/

I just tested with below Sample YAML file:

with 3 replicas:

scaled to 6 replicas:(Distributed evenly on all the nodes)

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx
spec:
  replicas: 3
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      topologySpreadConstraints:
        - maxSkew: 1
          topologyKey: kubernetes.io/hostname
          whenUnsatisfiable: ScheduleAnyway
          labelSelector:
            matchLabels:
              app: nginx
      containers:
      - name: nginx
        image: nginx