交付cadvisor

# 200机器，下载镜像：
~]# docker pull google/cadvisor:v0.28.3
~]# docker images|grep cadvisor
~]# docker tag 75f88e3ec33 harbor.od.com/public/cadvisor:v0.28.3
~]# docker push harbor.od.com/public/cadvisor:v0.28.3
~]# mkdir /data/k8s-yaml/cadvisor/
~]# cd /data/k8s-yaml/cadvisor/
cadvisor]# vi ds.yaml
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: cadvisor
  namespace: kube-system
  labels:
    app: cadvisor
spec:
  selector:
    matchLabels:
      name: cadvisor
  template:
    metadata:
      labels:
        name: cadvisor
    spec:
      hostNetwork: true
      tolerations:
      - key: node-role.kubernetes.io/master
        effect: NoExecute
      containers:
      - name: cadvisor
        image: harbor.od.com/public/cadvisor:v0.28.3
        imagePullPolicy: IfNotPresent
        volumeMounts:
        - name: rootfs
          mountPath: /rootfs
          readOnly: true
        - name: var-run
          mountPath: /var/run
        - name: sys
          mountPath: /sys
          readOnly: true
        - name: docker
          mountPath: /var/lib/docker
          readOnly: true
        ports:
          - name: http
            containerPort: 4194
            protocol: TCP
        readinessProbe:
          tcpSocket:
            port: 4194
          initialDelaySeconds: 5
          periodSeconds: 10
        args:
          - --housekeeping_interval=10s
          - --port=4194
      terminationGracePeriodSeconds: 30
      volumes:
      - name: rootfs
        hostPath:
          path: /
      - name: var-run
        hostPath:
          path: /var/run
      - name: sys
        hostPath:
          path: /sys
      - name: docker
        hostPath:
          path: /data/docker

此时我们看到大多数节点都运行在21机器上，我们人为的让pod调度到22机器（当然即使你的大多数节点都运行在22机器上也没关系） 可人为影响K8S调度策略的三种方法：

• 污点、容忍方法：
  • 污点：运算节点node上的污点（先在运算节点上打标签等 kubectl taint nodes node1 key1=value1:NoSchedule），污点可以有多个
  • 容忍度：pod是否能够容忍污点
  • 参考kubernetes官网
• nodeName：让Pod运行再指定的node上
• nodeSelector：通过标签选择器，让Pod运行再指定的一类node上

# 给21机器打个污点，22机器：
~]# kubectl taint node hdss7-21.host.com node-role.kubernetes.io/master=master:NoSchedule

# 21/22两个机器，修改软连接：
~]# mount -o remount,rw /sys/fs/cgroup/
~]# ln -s /sys/fs/cgroup/cpu,cpuacct /sys/fs/cgroup/cpuacct,cpu
~]# ls -l /sys/fs/cgroup/

mount -o remount, rw /sys/fs/cgroup ：重新以可读可写的方式挂载为已经挂载/sys/fs/cgroup

ln -s ：创建对应的软链接

ls -l ：显示不隐藏的文件与文件夹的详细信息

# 22机器，应用资源清单：
~]# kubectl apply -f http://k8s-yaml.od.com/cadvisor/ds.yaml
~]# kubectl get pods -n kube-system -o wide

# 21机器，我们删掉污点：
~]# kubectl taint node hdss7-21.host.com node-role.kubernetes.io/master-
# out: node/hdss7-21.host.com untainted