Pod Affinity
Pod 亲和性是指将新 pod 调度到已经拥有一个或多个满足特定条件的 pod 的节点上的能力。 Karpenter 通过为节点定义 pod 亲和性规则来支持 pod 亲和性。 在本节我们将了解如何使用 podAffinity 确保将前端 pod 部署在后端 pod 所在的同一区域中。 这里我们展示了 podAffinity 的示例，但它对于 podAntiAffinity 的工作原理也相同（允许将新的 pod 调度到不具有任何满足特定条件的现有 pod 的节点上）。
首先删除前面章节创建的资源：
kubectl delete deployment inflate
kubectl delete nodepools.karpenter.sh default
kubectl delete ec2nodeclasses.karpenter.k8s.aws default
部署NodePool：
mkdir ~/environment/karpenter
cd ~/environment/karpenter
cat <<EoF> podaffinity.yaml
apiVersion: karpenter.sh/v1beta1
kind: NodePool
metadata:
name: default
spec:
disruption:
consolidationPolicy: WhenUnderutilized
expireAfter: Never
limits:
cpu: "20"
template:
metadata:
labels:
eks-immersion-team: my-team
spec:
nodeClassRef:
name: default
# Requirements that constrain the parameters of provisioned nodes.
# These requirements are combined with pod.spec.affinity.nodeAffinity rules.
# Operators { In, NotIn } are supported to enable including or excluding values
requirements:
- key: "karpenter.k8s.aws/instance-category"
operator: In
values: ["c", "m", "r"]
- key: "kubernetes.io/arch"
operator: In
values: ["amd64"]
- key: "karpenter.sh/capacity-type" # If not included, the webhook for the AWS cloud provider will default to on-demand
operator: In
values: ["on-demand"]
---
apiVersion: karpenter.k8s.aws/v1beta1
kind: EC2NodeClass
metadata:
name: default
spec:
amiFamily: AL2
role: "KarpenterNodeRole-${CLUSTER_NAME}"
securityGroupSelectorTerms:
- tags:
alpha.eksctl.io/cluster-name: $CLUSTER_NAME
subnetSelectorTerms:
- tags:
alpha.eksctl.io/cluster-name: $CLUSTER_NAME
tags:
intent: apps
managed-by: karpenter
EoF
kubectl apply -f podaffinity.yaml
部署应用
export AZ1="$AWS_REGION"b
export AZ2="$AWS_REGION"c
我们将部署两个应用（backend和inflate）, backend应用设置了nodeAffinity，要部署在AZ2, inflate应用要部署在AZ1:
cd ~/environment/karpenter
cat <<EoF> nodeaffinity-pod-deploy.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: backend
spec:
replicas: 2
selector:
matchLabels:
app: backend
template:
metadata:
labels:
app: backend
spec:
affinity:
nodeAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
nodeSelectorTerms:
- matchExpressions:
- key: "topology.kubernetes.io/zone"
operator: "In"
values: ["$AZ2"]
terminationGracePeriodSeconds: 0
containers:
- name: backend
image: public.ecr.aws/eks-distro/kubernetes/pause:3.7
resources:
requests:
cpu: 1
nodeSelector:
eks-immersion-team: my-team
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: inflate
spec:
replicas: 2
selector:
matchLabels:
app: inflate
template:
metadata:
labels:
app: inflate
spec:
affinity:
nodeAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
nodeSelectorTerms:
- matchExpressions:
- key: "topology.kubernetes.io/zone"
operator: "In"
values: ["$AZ1"]
terminationGracePeriodSeconds: 0
containers:
- name: inflate
image: public.ecr.aws/eks-distro/kubernetes/pause:3.7
resources:
requests:
cpu: 1
nodeSelector:
eks-immersion-team: my-team
EoF
kubectl apply -f nodeaffinity-pod-deploy.yaml
这样Karpenter会在两个AZ各拉起一台机器：
且两台机器上各跑着两个pod：
kongpingfan:~/environment/karpenter $ kubectl get pod -o=custom-columns=NODE:.spec.nodeName,NAME:.metadata.name
NODE NAME
ip-192-168-130-42.us-west-2.compute.internal backend-7fb8544cf9-6n5mq
ip-192-168-130-42.us-west-2.compute.internal backend-7fb8544cf9-z547g
ip-192-168-189-213.us-west-2.compute.internal inflate-7c56688b5d-bwm76
ip-192-168-189-213.us-west-2.compute.internal inflate-7c56688b5d-cwmxz
测试podAffinity
现在部署frontend应用，frontend应用需要跟backend应用部署在同一个AZ(为了减少跨AZ访问流量)。所以我们为它添加podAffinity:
cd ~/environment/karpenter
cat <<EoF> podaffinity-deploy.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: frontend
spec:
replicas: 2
selector:
matchLabels:
app: frontend
template:
metadata:
labels:
app: frontend
spec:
affinity:
podAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchLabels:
app: backend
topologyKey: topology.kubernetes.io/zone
terminationGracePeriodSeconds: 0
containers:
- name: frontend
image: public.ecr.aws/eks-distro/kubernetes/pause:3.7
resources:
requests:
cpu: 1
nodeSelector:
eks-immersion-team: my-team
EoF
kubectl apply -f podaffinity-deploy.yaml
可以看到frontend的pod全部和backend在同一个AZ，并且分布在两个节点上(因为第一个节点的资源已经不够用了，要创建出来一个新的节点)：
kongpingfan:~/environment/karpenter $ kubectl get pod -o=custom-columns=NODE:.spec.nodeName,NAME:.metadata.name
NODE NAME
ip-192-168-130-42.us-west-2.compute.internal backend-7fb8544cf9-6n5mq
ip-192-168-130-42.us-west-2.compute.internal backend-7fb8544cf9-z547g
ip-192-168-36-213.us-west-2.compute.internal frontend-8547476cdc-krdtn
ip-192-168-130-42.us-west-2.compute.internal frontend-8547476cdc-tv59n
ip-192-168-189-213.us-west-2.compute.internal inflate-7c56688b5d-bwm76
ip-192-168-189-213.us-west-2.compute.internal inflate-7c56688b5d-cwmxz