Schedule pods based on inter-pod affinity - Container Compute Service

The inter-pod affinity scheduling policy is used to express affinity preferences between pods. Compared with node affinity, inter-pod affinity scheduling limits the nodes to which pods can be scheduled based on the labels of pods that already run on nodes. In an Alibaba Cloud Container Compute Service (ACS) cluster, you can use Kubernetes-native scheduling semantics to implement inter-pod affinity scheduling. By specifying the topology domain and label rules in the podAffinity or podAntiAffinity field, you can schedule pods to the specified topology domain. This topic describes the limits and usage notes of inter-pod affinity scheduling in ACS.

Prerequisites

An ACS cluster is created. For more information, see Create an ACS cluster.
kube-scheduler is installed. For more information, see kube-scheduler.
acs-virtual-node v2.12.0-acs.4 or later is installed.

Usage notes

Inter-pod affinity contains the affinity and anti-affinity modes. Both modes have the same protocol format and consist of the requiredDuringSchedulingIgnoredDuringExecution rule and the preferredDuringSchedulingIgnoredDuringExecution rule. The TopologyKey field is required, which corresponds to the label of the virtual node. The following table describes the topology labels supported by ACS for different types of nodes.

Node type	Label	Description	Example
Regular node	topology.kubernetes.io/zone	Network zone	topology.kubernetes.io/zone: cn-shanghai-b

ACS supports multiple compute classes. The following constraints exist in different compute classes when you use other fields of topology distribution constraints.

requiredDuringSchedulingIgnoredDuringExecution

Compute class	Field	Description	Constraint
General-purpose Performance-enhanced	LabelSelector	This field is used to find matching pods. Pods that match this label selector are counted to determine the number of pods in the topology domain.	Pods in other compute classes, such as the GPU-accelerated and performance-enhanced compute classes, are not counted.
	Namespaces	This field is used to find matching namespaces and can be used together with the LabelSelector field.
	NamespaceSelector	This field is similar to the Namespaces field, except it retrieves namespaces based on the namespace label.
GPU-accelerated	LabelSelector	This field is used to find matching pods. Pods that match this label selector are counted to determine the number of pods in the topology domain.	Pods in other compute classes, such as the general-purpose and performance-enhanced compute classes, are not counted.
	Namespaces	This field is used to find matching namespaces and can be used together with the LabelSelector field.	Not supported
	NamespaceSelector	This field is similar to the Namespaces field, except it retrieves namespaces based on the namespace label.	Not supported

preferredDuringSchedulingIgnoredDuringExecution

Compute class

Field

Description

Constraint

General-purpose
Performance-enhanced

LabelSelector

Namespaces

NamespaceSelector

None

Pods in other compute classes, such as the GPU-accelerated and performance-enhanced compute classes, are not counted.

For more information about the fields, see Inter-pod affinity and anti-affinity.

Examples

The following example shows how to schedule pods to a specific zone by configuring the podAffinity field.

Run the following command to view the nodes in the cluster:

kubectl get node

Expected output:

NAME                            STATUS   ROLES   AGE     VERSION
virtual-kubelet-cn-hangzhou-i   Ready    agent   5h42m   v1.28.3-xx
virtual-kubelet-cn-hangzhou-j   Ready    agent   5h42m   v1.28.3-xx

Create a file named with-affinity-pod.yaml and add the following content to the file:

apiVersion: v1
kind: Pod
metadata:
  labels:
    pod-affinity-label: with-pod-affinity
  name: with-affinity-label-pod
spec:
  containers:
    - args:
        - 'infinity'
      command:
        - sleep
      image: registry-cn-hangzhou.ack.aliyuncs.com/acs/stress:v1.0.4
      imagePullPolicy: IfNotPresent
      name: stress
      resources:
        limits:
          cpu: '1'
          memory: 1Gi
        requests:
          cpu: '1'
          memory: 1Gi

Run the following command to deploy the with-affinity-label-pod file in the cluster:
```
kubectl apply -f with-affinity-pod.yaml
```

Run the following command to view the distribution results of pods:

kubectl get pod -o wide

Expected output:

NAME                      READY   STATUS    RESTARTS   AGE   IP              NODE                            NOMINATED NODE   READINESS GATES
with-affinity-label-pod   1/1     Running   0          75s   192.168.xx.xxx  virtual-kubelet-cn-hangzhou-i   <none>           <none>

The output indicates that the pods are scheduled to the cn-hangzhou-i zone.

Create a file named origin-affinity-pod.yaml and add the following content to the file:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: dep-pod-affinity
  labels:
    app: pod-affinity-demo
spec:
  replicas: 4
  selector:
    matchLabels:
      app: pod-affinity-demo
  template:
    metadata:
      labels:
        app: pod-affinity-demo
    spec:
      containers:
      - name: pod-affinity-demo
        image: registry-cn-hangzhou.ack.aliyuncs.com/acs/stress:v1.0.4
        command:
        - "sleep"
        - "infinity"
        resources:
          limits:
            cpu: '1'
            memory: 1Gi
          requests:
            cpu: '1'
            memory: 1Gi
      # Specify the affinity between pods. This pod must be deployed in the same zone as the pod with the <pod-affinity-label:with-pod-affinity> label.
      affinity:
        podAffinity:
          requiredDuringSchedulingIgnoredDuringExecution:
          - labelSelector:
              matchExpressions:
              - key: pod-affinity-label
                operator: In
                values:
                - with-pod-affinity
            topologyKey: topology.kubernetes.io/zone

Run the following command to deploy the origin-affinity-pod.yaml file in the cluster:
```
kubectl apply -f origin-affinity-pod.yaml
```

Run the following command to view the distribution results of pods:

kubectl get pod -o wide

Expected output:

NAME                                READY   STATUS    RESTARTS   AGE     IP              NODE                            NOMINATED NODE   READINESS GATES
dep-pod-affinity-6b9d4f7c87-5jlfx   1/1     Running   0          3m26s   192.168.xx.xxx  virtual-kubelet-cn-hangzhou-i   <none>           <none>
dep-pod-affinity-6b9d4f7c87-hwdpc   1/1     Running   0          3m26s   192.168.xx.xxx  virtual-kubelet-cn-hangzhou-i   <none>           <none>
dep-pod-affinity-6b9d4f7c87-jfcrq   1/1     Running   0          3m26s   192.168.xx.xxx  virtual-kubelet-cn-hangzhou-i   <none>           <none>
dep-pod-affinity-6b9d4f7c87-xwbfr   1/1     Running   0          3m26s   192.168.xx.xxx  virtual-kubelet-cn-hangzhou-i   <none>           <none>
with-affinity-label-pod             1/1     Running   0          6m30s   192.168.xx.xxx  virtual-kubelet-cn-hangzhou-i   <none>           <none>

The output indicates that all pods are distributed in the same zone.