k8s中的多pod节点动态控制资源倾斜给其中某一个pod

以nginx-test-limit1，nginx-test-limit2为例
1，查出容器id
kubectl describe pod nginx-test-limit1

2，容器id：3b4a485c589176227157cf394e31f8d4962a7700d1585f58002356ce9affb061

2，通过容器id查到Cgroup的路径：
docker inspect 3b4a485c5891|grep Cgroup

3，进入容器的cpu配置文件目录：
/sys/fs/cgroup/cpu/kubepods.slice/kubepods-besteffort.slice/kubepods-besteffort-pod6fb1f21d_c7ab_4d2f_a621_daea4f8aa113.slice

4，cpu配额文件cpu.cfs_quota_us，默认为-1

5，将cpu.cfs_quota_us修改成30%cpu配额，不需要重启pod：
echo 30000 > cpu.cfs_quota_us

此时nginx-test-limit1的cpu使用率不超过30%
nginx-test-limit2的cpu使用率在150%左右

6，将nginx-test-limit的cpu配额文件cpu.cfs_quota_us，改回为-1，不需要重启pod：
echo -1 > cpu.cfs_quota_us

此时nginx-test-limit1资源在50%-120%之间浮动
nginx-test-limit2资源同样在50%-120%之间浮动

7，提高cpu使用率脚本：
#! /bin/sh
stop()
{
echo “”
}
start()
{
for i in seq $1
do
echo "
i=0;
while true
do
i=i+1;
done" | /bin/sh &
echo $! >> pid.txt
done
}
case $1 in
stop)
stop
;;
*)
start $1
;;
Esac

8，pod编排文件：
1，nginx-test-limit1：
apiVersion: v1
kind: Pod
metadata:
name: nginx-test-limit1
annotations:
lwt1: add
labels:
app: nginx-test-limit1
spec:
nodeName: moss-node48
containers:

• name: nginx-test-limit1
  image: harbor.com/library/nginx:1.23
  command: [“/bin/sleep”,“infinity”]
  imagePullPolicy: IfNotPresent
  #resources:

  limits:

  #cpu: "3000m"
  #memory: "100Mi"
  

  #requests:
  #cpu: “1000m”
  #memory: “100Mi”
  ports:
  - containerPort: 180
  priorityClassName: high-priority-nonpreempting

2，nginx-test-limit2
apiVersion: v1
kind: Pod
metadata:
name: nginx-test-limit2
annotations:
lwt1: add
labels:
app: nginx-test-limit2
spec:
nodeName: moss-node48
containers:

• name: nginx-test-limit2
  image: harbor.com/library/nginx:1.23
  command: [“/bin/sleep”,“infinity”]
  imagePullPolicy: IfNotPresent
  ports:
  - containerPort: 181
  priorityClassName: second-nonpreempting