在《研发工程师玩转Kubernetes——利用Pod反亲和性控制一个Node上只能有一个Pod》一文中我们介绍了Pod的反亲和性应用。本节我们将使用亲和性做几个实验。
nginx Pod在每个Node上只能存在一个；alpine Pod在每个Node上只能存在一个，同时要求该Node上还要有nginx Pod。
实验分为3组：
- 创建2个nginx Pod和2个alpine Pod，观察它们在Node上的分布。
- 将alpine Pod数调整到3，观察其状态；再将nginx Pod数调整到3，观察整体变化。
- 将nginx Pod数调整到2，观察整体变化。
清单文件
nginx清单文件
代码语言：javascript复制# nginx_deployment_pod_affinity.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: nginx-deployment
spec:
replicas: 2
selector:
matchLabels:
app: nginx
template:
metadata:
labels:
app: nginx
spec:
affinity:
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: app
operator: In
values:
- nginx
topologyKey: "kubernetes.io/hostname"
containers:
- name: nginx-container
image: nginx
ports:
- containerPort: 80该清单文件指定一个Node上只能有一个nginx Pod。
alpine清单文件
代码语言：javascript复制# alpine_deployment_pod_affinity.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: alpine-deployment
spec:
replicas: 2
selector:
matchLabels:
app: alpine
template:
metadata:
labels:
app: alpine
spec:
affinity:
podAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: app
operator: In
values:
- nginx
topologyKey: "kubernetes.io/hostname"
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: app
operator: In
values:
- alpine
topologyKey: "kubernetes.io/hostname"
containers:
- name: alpine-container
image: alpine
stdin: true
tty: true该清单文件指定在每个Node上只能存在一个alpine Pod，同时要求该Node上还要有nginx Pod。
实验
创建2个nginx Pod和2个alpine Pod
代码语言：javascript复制kubectl create -f nginx_deployment_pod_affinity.yaml -f alpine_deployment_pod_affinity.yaml代码语言：javascript复制deployment.apps/nginx-deployment created deployment.apps/alpine-deployment created
kubectl get pod -o wideNAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
nginx-deployment-57555788b6-t8xqg 1/1 Running 0 8s 10.1.226.8 ubuntue <none> <none>
nginx-deployment-57555788b6-4c58f 1/1 Running 0 8s 10.1.43.214 ubuntuc <none> <none>
alpine-deployment-7f8b7cf975-vxvkq 1/1 Running 0 8s 10.1.226.9 ubuntue <none> <none>
alpine-deployment-7f8b7cf975-cfvqm 1/1 Running 0 8s 10.1.43.215 ubuntuc <none> <none>可以看到，alpine按照我们的预期被调度到存在nginx的Pod上。
将alpine Pod数调整到3，再将nginx Pod数调整到3
将alpine Pod数调整到3
代码语言：javascript复制kubectl scale --replicas=3 deployment alpine-deployment deployment.apps/alpine-deployment scaled
使用下面指令观察Pod的变化
代码语言：javascript复制kubectl get pod --watchNAME READY STATUS RESTARTS AGE
nginx-deployment-57555788b6-t8xqg 1/1 Running 0 8m44s
nginx-deployment-57555788b6-4c58f 1/1 Running 0 8m44s
alpine-deployment-7f8b7cf975-vxvkq 1/1 Running 0 8m44s
alpine-deployment-7f8b7cf975-cfvqm 1/1 Running 0 8m44s
alpine-deployment-7f8b7cf975-jjzqc 0/1 Pending 0 1s可以看到alpine一直处于Pending状态。因为它要求被调度的Node上有nginx Pod，而又不能有alpine Pod。这个也符合预期。
在新窗口中，可以通过下面指令看到其状态
代码语言：javascript复制kubectl describe pod alpine-deployment-7f8b7cf975-jjzqc Name: alpine-deployment-7f8b7cf975-jjzqc
Namespace: default
Priority: 0
Service Account: default
Node: <none>
Labels: app=alpine
pod-template-hash=7f8b7cf975
Annotations: <none>
Status: Pending
IP:
IPs: <none>
Controlled By: ReplicaSet/alpine-deployment-7f8b7cf975
Containers:
alpine-container:
Image: alpine
Port: <none>
Host Port: <none>
Environment: <none>
Mounts:
/var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-rqrls (ro)
Conditions:
Type Status
PodScheduled False
Volumes:
kube-api-access-rqrls:
Type: Projected (a volume that contains injected data from multiple sources)
TokenExpirationSeconds: 3607
ConfigMapName: kube-root-ca.crt
ConfigMapOptional: <nil>
DownwardAPI: true
QoS Class: BestEffort
Node-Selectors: <none>
Tolerations: node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning FailedScheduling 4m38s default-scheduler 0/5 nodes are available: 2 node(s) didn't match pod anti-affinity rules, 3 node(s) didn't match pod affinity rules. preemption: 0/5 nodes are available: 2 No preemption victims found for incoming pod, 3 Preemption is not helpful for scheduling..将nginx Pod数调整到3
在新窗口中执行下面指令调整
代码语言：javascript复制kubectl scale --replicas=3 deployment nginx-deployment 再回到之前观察Pod变化的窗口
代码语言：javascript复制NAME READY STATUS RESTARTS AGE
nginx-deployment-57555788b6-t8xqg 1/1 Running 0 8m44s
nginx-deployment-57555788b6-4c58f 1/1 Running 0 8m44s
alpine-deployment-7f8b7cf975-vxvkq 1/1 Running 0 8m44s
alpine-deployment-7f8b7cf975-cfvqm 1/1 Running 0 8m44s
alpine-deployment-7f8b7cf975-jjzqc 0/1 Pending 0 1s
nginx-deployment-57555788b6-hmklk 0/1 Pending 0 0s
nginx-deployment-57555788b6-hmklk 0/1 Pending 0 0s
alpine-deployment-7f8b7cf975-jjzqc 0/1 Pending 0 6m46s
nginx-deployment-57555788b6-hmklk 0/1 ContainerCreating 0 0s
alpine-deployment-7f8b7cf975-jjzqc 0/1 ContainerCreating 0 6m46s
nginx-deployment-57555788b6-hmklk 0/1 ContainerCreating 0 0s
alpine-deployment-7f8b7cf975-jjzqc 0/1 ContainerCreating 0 6m46s
nginx-deployment-57555788b6-hmklk 1/1 Running 0 4s
alpine-deployment-7f8b7cf975-jjzqc 1/1 Running 0 6m52s可以看到处于Pending状态的alpine在新扩出来的nginx Pod还处在ContainerCreating时，也变成了ContainerCreating状态，并在之后变成running状态。
将nginx Pod数调整到2
代码语言：javascript复制kubectl scale --replicas=2 deployment nginx-deploymentdeployment.apps/nginx-deployment scaled
观察整体变化
代码语言：javascript复制kubectl get pod -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
nginx-deployment-57555788b6-t8xqg 1/1 Running 0 20m 10.1.226.8 ubuntue <none> <none>
nginx-deployment-57555788b6-4c58f 1/1 Running 0 20m 10.1.43.214 ubuntuc <none> <none>
alpine-deployment-7f8b7cf975-vxvkq 1/1 Running 0 20m 10.1.226.9 ubuntue <none> <none>
alpine-deployment-7f8b7cf975-cfvqm 1/1 Running 0 20m 10.1.43.215 ubuntuc <none> <none>
alpine-deployment-7f8b7cf975-jjzqc 1/1 Running 0 11m 10.1.202.201 ubuntud <none> <none>可以看到alpine并没有变化。因为亲和性只是调度，所以在亲和条件发生变化时，之前已经调度好的结果并不会改变。
