k8s学习（21）：调度亲和性

发表于 更新于 分类于

节点亲和性

pod.spec.nodeAffinity

  • preferredDuringSchedulingIgnoredDuringExecution：软策略 (百度翻译：优先执行计划)
  • requiredDuringSchedulingIgnoredDuringExecution：硬策略（百度翻译：执行期间要求的预定计划）

老师讲解：

  • 软策略：有更优的节点就选择更优的节点
  • 硬策略：必须选择我指定的将节点

requiredDuringSchedulingIgnoredDuringExecution

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
apiVersion:  v1
kind: Pod
metadata:
  name: affinity
  labels:
    app: node-affinity-pod
spec:
  containers:
  - name: with-node-affinity
    image: hub.test.com/library/myapp:v1
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: kubernetes.io/hostname  #key就是节点的标签，使用`kubectl get node --show-labels`
              operator: NotIn
              values:
              - k8s-node02     ## node节点不为node02

键值运算关系

  • In：label的值在某个列表中
  • NotIn：label的值不在某个列表中
  • Gt：label的值大于某个值
  • Lt：label的值小于某个值
  • Exists：某个label存在
  • DoesNotExist：某个label不存在
1
2
3
4
5
6
# 创建
kubectl create -f pod.yaml
# 查看
[root@k8s-master affinity]# kubectl get pods -o wide
NAME           READY   STATUS    RESTARTS   AGE   IP             NODE         NOMINATED NODE   READINESS GATES
affinity-pod   1/1     Running   0          45s   10.244.1.117   k8s-node01   <none>           <none>

PreferredDuringSchedulingIgnoredDuringExectuion

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
apiVersion: v1
kind: Pod
metadata:
  name: affinity-preferred
  labels:
    app: node-affinity-pod
spec:
  containers:
  - name: with-node-affinity
    image: hub.test.com/library/myapp:v1
  affinity:
    nodeAffinity:
      preferredDuringSchedulingIgnoredDuringExecution:
      - weight: 1			## 权重，如果有多个，则优先选择权重高的
        preference:
          matchExpressions:
            - key: kubernetes.io/hostname
              operator: In
              values:
              - k8s-node03
1
2
3
4
5
6
7
[root@k8s-master affinity]# kubectl create -f pod2.yaml 
pod/affinity-preferred created
[root@k8s-master affinity]# kubectl get pods -o wide
NAME                 READY   STATUS    RESTARTS   AGE   IP             NODE         NOMINATED NODE   READINESS GATES
affinity-pod         1/1     Running   0          19m   10.244.1.117   k8s-node01   <none>           <none>
affinity-preferred   1/1     Running   0          7s    10.244.2.208   k8s-node02   <none>           <none>

同时使用软策略和硬策略

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
apiVersion: v
kind: Pod
metadata:
  name: faainity-multi
  app: node-affinity-pod
spec:
  containers:
  - name: with-node-affinity
    image: hub.test.com/library/myapp:v1
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoreDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: kubernetes.io/hostname
            operator: NotIn
            values:
            - k8s-node02
      preferredDuringSchedulingIgnoreDuringExecution:
      - weight: 1
        preference:
        matchExpressions:
        - key: source
          operator: In
          - qikaiak

Pod亲和性

Pod.spec.affinity.podAffinity/podAntiAffinity 亲和性/反亲和性

preferredDuringSchedulingIgnoredDuringExection：软策略

requiredDruingSchedulingIgnoredDuringExecution：硬策略

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
apiVersion: v1
kind: Pod
metadata:
  name: pod-3
  labels:
    app: pod-3
spec:
  containers:
  - name: pod-3
    image: hub.test.com/library/myapp:v1
  affinity:
    podAffinity:
      requiredDuringSchdeulingIgnoredDuringExecution:
      - labelSelector:
          matchExpressions:
          - key: app
            operator: In
            values:
            - pod-1
        topologyKey: kubernetes.io/hostname    ## 拓扑域
    podAntiAffinity:
      preferredDuringSchedulingIgnoredDuringExecution:
      - weight: 1
        podAffinityTerm:
          labelSelector:
            matchExpressions:
            - key: app
              operator: In
              values:
              - pod-2
          topologyKey: kubernetes.io/hostname

这里的podAffinityTerm和nodeSelectorTerms，一个有s，一个没有

1
2
3
4
5
6
7
8
9
10
11
12
[root@k8s-master affinity]# kubectl create -f pod3.yaml 
pod/affinity-pod-3 created
[root@k8s-master affinity]# kubectl get pods
NAME                 READY   STATUS    RESTARTS   AGE
affinity-pod         1/1     Running   0          115m
affinity-pod-3       0/1     Pending   0          5s
affinity-preferred   1/1     Running   0          96m
[root@k8s-master affinity]# kubectl get pods -o wide --show-labels
NAME                 READY   STATUS    RESTARTS   AGE    IP             NODE         NOMINATED NODE   READINESS GATES   LABELS
affinity-pod         1/1     Running   0          116m   10.244.1.117   k8s-node01   <none>           <none>            app=node-affinity-pod
affinity-pod-3       0/1     Pending   0          18s    <none>         <none>       <none>           <none>            <none>
affinity-preferred   1/1     Running   0          96m    10.244.2.208   k8s-node02   <none>           <none>            <none>

可以看到pod-3是pending状态

修改其中一个pod的标签app=pod-1

1
2
3
4
5
6
7
[root@k8s-master affinity]# kubectl label pod affinity-pod app=pod-1 --overwrite=true
pod/affinity-pod labeled
[root@k8s-master affinity]# kubectl get pods -o wide --show-labels
NAME                 READY   STATUS    RESTARTS   AGE    IP             NODE         NOMINATED NODE   READINESS GATES   LABELS
affinity-pod         1/1     Running   0          136m   10.244.1.117   k8s-node01   <none>           <none>            app=pod-1
affinity-pod-3       1/1     Running   0          21m    10.244.1.118   k8s-node01   <none>           <none>            <none>
affinity-preferred   1/1     Running   0          117m   10.244.2.208   k8s-node02   <none>           <none>            <none>

此时pod-3的状态就变成Running状态

这里的含义就是：判断在kubernetes.io/hostname下，app必须包含pod-1，可选不包含pod-2

亲和性/反亲和性调度策略比较如下：

image-20200204211100740