kubernetes之pod调度

调度规则

deployment全自动调度： 运行在哪个节点上完全由master的scheduler经过一系列的算法计算得出, 用户无法进行干预
nodeselector定向调度： 指定pod调度到一些node上, 通过设置node的标签和deployment的nodeSelctor属性相匹配
	多个node有相同标签, scheduler 会选择一个可用的node进行调度
	nodeselector定义的标签匹配不上任何node上的标签， 这个pod是无法调度
	k8s给node预定义了一些标签， 通过kubectl describe node xxxx进行查看
	用户可以使用k8s给node预定义的标签
NodeAffinity： node节点亲和性
	硬限制 ： 必须满足指定的规则才可以调度pod到node上
	软限制 ： 优先满足指定规则，调度器会尝试调度pod到node上， 但不强求
	
PodAffinity:  pod亲和与互斥调度， 根据在节点上正在运行的pod标签而不是节点的标签进行判断和调度
	亲和： 匹配标签两个pod调度到同一个node上
	互斥： 匹配标签两个pod不能运行在同一个node上

nodeselector定向调度

给目标node打上一些标签

kubectl label nodes  <node-name> <label-key>=<label-value>

kubectl label nodes 192.168.1.62 disk=ssd

在pod的定义加上nodeSelector设置

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
spec:
  selector:
    matchLabels:
      app: nginx
  replicas: 1
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx:1.11
        ports:
        - containerPort: 80
      nodeSelector:
        disk: ssd

如果给多个node都定义了相同的标签, 则调度器会根据调度算法从这组node中挑选一个可用的node进行pod调度
如果指定了pod的nodeSelector条件, 且集群中不存在包含响应标签的Node, 则即使在集群中还有其它可供使用的Node,这个pod也无法被成功调度
用户可以自行给node添加标签, kubernetes也会给node预定义一些标签, 可以通过kubectl describe node

Labels:             beta.kubernetes.io/arch=amd64
                    beta.kubernetes.io/os=linux
                    disk=ssd
                    kubernetes.io/hostname=192.168.1.62

uploading-image-951691.png

NodeAffinity

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
spec:
  selector:
    matchLabels:
      app: nginx
  replicas: 4
  template:
    metadata:
      labels:
        app: nginx
    spec:
      affinity:
        nodeAffinity:
          requiredDuringSchedulingIgnoredDuringExecution:
              nodeSelectorTerms:
              - matchExpressions:
                - key: beta.kubernetes.io/arch
                  operator: In
                  values:
                  - amd64
          preferredDuringSchedulingIgnoredDuringExecution:
          - weight: 1
            preference:
              matchExpressions:
              - key: disk
                operator: In
                values:
                - ssd
      containers:
      - name: nginx
        image: nginx:1.11
        ports:
        - containerPort: 80

requiredDuringSchedulingIgnoredDuringExecution: 硬限制, 要求只运行在amd64的节点上
preferredDuringSchedulingIgnoredDuringExecution: 软限制, 尽量运行在磁盘类型ssd的节点上

PodAffinity: pod亲和性调度

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment-test1
spec:
  selector:
    matchLabels:
      app: nginx-test1
  replicas: 1
  template:
    metadata:
      labels:
        app: nginx-test1
        security: s1        
    spec:
      containers:
      - name: nginx
        image: nginx:1.11
        ports:
        - containerPort: 80

---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment-test2
spec:
  selector:
    matchLabels:
      app: nginx-test2
  replicas: 4
  template:
    metadata:
      labels:
        app: nginx-test2
        security: s1
    spec:
      affinity:
        podAffinity:
          requiredDuringSchedulingIgnoredDuringExecution:
          - labelSelector:
              matchExpressions:
              - key: app
                operator: In
                values:
                - nginx-test1
            topologyKey: kubernetes.io/hostname     
      containers:
      - name: nginx
        image: nginx:1.11
        ports:
        - containerPort: 80