Kubernetes 为 Namespace 配置pod、CPU和内存配额

Kubernetes 为 Namespace 配置pod、CPU和内存配额

Kubernetes 为 Namespace 配置POD、CPU和内存配额

• • 一、介绍
  • POD配额实验：
  • 三、CPU、内存配额实验

一、介绍

前面namespace中说到，Namespace是一种将集群资源划分为多个用途(通过 resource quota)的方法。在k8s上准入控制器的模块有很多，其中比较常用的有LimitRanger、ResourceQuota、ServiceAccount、PodSecurityPolicy等等，对于前面三种准入控制器系统默认是启用的，我们只需要定义对应的规则即可

POD配额实验：

1.创建namespace.yaml文件创建namespace

apiVersion: v1
kind: Namespace
metadata:
  name: resourcequota

[root@iZbp16ke9g4qwtz7gw25fnZ ~]# kubectl create -f namespace.yaml 
namespace/resourcequota created
[root@iZbp16ke9g4qwtz7gw25fnZ ~]# kubectl get ns | grep resourcequota
resourcequota      Active   17s

2.创建resourcequota为namespace设置pod数量限制

apiVersion: v1
kind: ResourceQuota
metadata:
  name: pod-quota1
  namespace: resourcequota
spec:
  hard:
    pods: 2

spec:hard即资源配额列表，其各字段的含义和单位如表所示。

[root@iZbp16ke9g4qwtz7gw25fnZ ~]# kubectl create -f quota-pod.yaml 
resourcequota/pod-quota1 created
[root@iZbp16ke9g4qwtz7gw25fnZ ~]# kubectl get resourcequota -n=resourcequota
NAME         CREATED AT
pod-quota1   2021-12-16T01:51:39Z

3.创建deployment(3pod)

apiVersion: apps/v1beta1
kind: Deployment
metadata:
  name: nginx-v1
  namespace: resourcequota
spec:
  replicas: 3
  template:
    metadata:
      labels:
        app: nginx-v1
    spec:
      containers:
      - name: nginx-quota
        image: hlqlinux/hlq-test:nginx-v3

由于做了pod资源限制，只有两个pod被创建

[root@iZbp16ke9g4qwtz7gw25fnZ ~]# kubectl get deploy -n=resourcequota
NAME       READY   UP-TO-DATE   AVAILABLE   AGE
nginx-v1   2/3     2            2           2m4s

限制成功

三、CPU、内存配额实验

本实验展示了如何为某一名字空间内运行的所有容器配置CPU和内存配额。配额可以通过 ResourceQuota对象设置。

报错： 0/2 nodes are available: 1 Insufficient cpu, 1 node(s) had taints that the pod didn’t tolerate.
设置请求的POD数为3,node节点cpu资源为1，导致节点调度不上去。

resourcequota资源请求配置

apiVersion: v1
kind: ResourceQuota
metadata:
  name: pod-quota1
  namespace: resourcequota
spec:
  hard:
    requests.cpu: 2
    requests.memory: 2Gi
    limits.cpu: 2
    limits.cpu: 2Gi

测试deployment

apiVersion: apps/v1beta1
kind: Deployment
metadata:
  name: nginx-v1
  namespace: resourcequota
spec:
  replicas: 3
  template:
    metadata:
      labels:
        app: nginx-v1
    spec: 
      containers:
      - name: nginx-quota
        image: hlqlinux/hlq-test:nginx-v3
        resources:
          limits:
            cpu: 0.2
            memory: 0.2Gi
          requests:
            cpu: 0.2
            memory: 0.2Gi

kubectl describe pod nginx-v1-7759dc485b-4g8hf -nresourcequota 
Containers:
  nginx-quota:
    Limits:
      cpu:     200m
      memory:  214748364800m
    Requests:
      cpu:        200m
      memory:     214748364800m