七点半的菜市场

large picture of the cover

十四、Kubernetes 资源限制

🏆 k8s容器资源限制

⭐️ kubern有两种限制类型来进行资源分配

1.request(资源需求)：运行Pod的节点必须满足运行Pod的最基本需求才能运行Pod

2,limit：运行Pod期间，可能内存使用量会增加，最多能使用多少内存，这就是资源限额。（2）资源类型：

CPU 的单位是核心数，内存的单位是字节

一个容器申请0.5个CPU，就相当于申请1个CPU的一半，你也可以加个后缀m 表示千分之一的概念。100m的CPU=0.1个CPU

内存单位：

K、M、G、T、P、E #通常是以1000为换算标准的。

Ki、Mi、Gi、Ti、Pi、Ei #通常是以1024为换算标准的

⭐️ 容器资源限制的类型

🍪 内存

1️⃣ stress是压力测试的镜像

[root@master resources]# cat pod-ram.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: memory-demo
spec:
  containers:
  - name: memory-demo
    image: joedval/stress:latest
    args:
    - --vm ##1vm默认=250mb
    - "1" 
    - --vm-bytes ##修改vm的内存为256mb
    - 200M   
    resources:
      requests:
        memory: 50Mi   ##最小内存
      limits:
        memory: 100Mi ##最大内存

-n, --dry-run show what would have been done // 指定运行多少秒

-t, --timeout N timeout after N seconds // 等待xx微秒后才开始运行

--backoff N    wait factor of N microseconds before work starts               // 产生多个处理sqrt()函数的CPU进程

-c, --cpu N spawn N workers spinning on sqrt() // 产生多个处理sync()函数的磁盘I/O进程

-i, --io N spawn N workers spinning on sync()

-m, --vm N spawn N workers spinning on malloc()/free()

--vm-bytes B   malloc B bytes per vm worker (default is 256MB)

--vm-stride B  touch a byte every B bytes (default is 4096)

--vm-hang N    sleep N secs before free (default none, 0 is inf)

--vm-keep      redirty memory instead of freeing and reallocating

-d, --hdd N spawn N workers spinning on write()/unlink() --hdd-bytes B write B bytes per hdd worker (default is 1GB)

2️⃣ 查看 pod

[root@master resources]# kubectl get pod
memory-demo                   0/1     CrashLoopBackOff   8          6m26s
[root@master resources]# kubectl describe pod memory-demo
    State:          Waiting
      Reason:       CrashLoopBackOff
    Last State:     Terminated
      Reason:       OOMKilled ##内存溢出

2️⃣ 修改内存限制

[root@master resources]# kubectl delete pod memory-demo ##删除旧pod
[root@master resources]# cat pod-ram.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: memory-demo
spec:
  containers:
  - name: memory-demo
    image: joedval/stress:latest
    args:
    - --vm
    - "1" 
    - --vm-bytes
    - 90M   
    resources:
      requests:
        memory: 50Mi  
      limits:
        memory: 100Mi 
    
[root@master resources]# kubectl apply -f pod-ram.yaml 
pod/memory-demo created
##启动成功
[root@master resources]# kubectl get pod
memory-demo                   1/1     Running   0          57s

🍪 Cpu

1️⃣ 编写 yaml文件

[root@master resources]# cat pod-cpu.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: cpu-pod
spec:
  containers:
  - image: joedval/stress:latest
    name: cpu-pod
    args:
    - c
    - "1"
    resources:
      requests:
        cpu: "1"
      limits:
        cpu: "1"

2️⃣ 执行

[root@master resources]# kubectl apply -f pod-cpu.yaml 
pod/cpu-pod created
##资源够可以启动
[root@master resources]# kubectl get pod
cpu-pod                   1/1     Running             0          8m16s

3️⃣ 修改资源配置

[root@master resources]# cat pod-cpu.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: cpu-pod
spec:
  containers:
  - image: joedval/stress
    name: cpu-pod
    args:
    - -c
    - "1"
    resources:
      requests:
        cpu: "500M"
      limits:
        cpu: "900M"

4️⃣ 启动查看，他就会一直调度，产生调度失败是因为申请的CPU资源超出集群节点所能提供的资源，但CPU 使用率过高，不会被杀死

[root@master resources]# kubectl get pod
cpu-pod                       0/1     Pending   0          7s

⭐️ 为namespace设置资源限制

LimitRange 在 namespace 中施加的最小和最大内存限制只有在创建和更新 Pod 时才会被应用。改变 LimitRange 不会对之前创建的 Pod 造成影响

[root@master resources]# cat pod-limitrange.yml 
apiVersion: v1
kind: LimitRange
metadata:
  name: limit-namespece
spec:
  limits:
  - default:  ##默认的最大配额，就是如果你创建pod没有指定pod那就是这个pod
      cpu: 1
      memory: 300Mi
    defaultRequest: ##同上，最小，这个值不能超过min
      cpu: 0.5  
      memory: 100Mi
    max:   ##假如你在创建pod的时候使用配额最大就只能是这个配额
      memory: 300Mi
      cpu: 2
    min: ##同上最小
      memory: 100Mi
      cpu: 0.5
    type: Container ##适合配额的类型

1️⃣ 创建

[root@master resources]# kubectl apply -f pod-limitrange.yml 
limitrange/limit-namespece created
##查看limitrange
[root@master resources]# kubectl get  limits
NAME              CREATED AT
limit-namespece   2021-09-16T12:18:35Z

2️⃣ 创建配置 pod

[root@master resources]# cat pod-cpu.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: cpu-pod
spec:
  containers:
  - image: joedval/stress
    name: cpu-pod
    args:
    - -c
    - "2"
    resources:
      requests:
        cpu: "2" #最小
      limits:
        cpu: "3"
    
##可以看到失败了，即使他在yml配置的配额内
[root@master resources]# kubectl apply -f pod-cpu.yaml 
Error from server (Forbidden): error when creating "pod-cpu.yaml": pods "cpu-pod" is forbidden: maximum cpu usage per Container is 2, but limit is 3

3️⃣ 为 namspace设置 reuquset限制

[root@master resources]# cat pod-resourcequot.yaml 
apiVersion: v1
kind: ResourceQuota
metadata:
  name: pod-resourcequota
spec:
  hard:
    requests.cpu: "1" ##至少有一个CPU
    requests.memory: "500Mi" ##至少有500M内存
    limits.cpu: "2" ##最多2个cpu
etadata:
  name: pod-resourcequota
spec:
  hard:
    requests.cpu: "1" ##至少有一个CPU
    requests.memory: "500Mi" ##至少有500M内存
    limits.cpu: "2" ##最多2个cpu
    limits.memory: "1Gi" ##最多1g内存

[root@master resources]# kubectl delete -f pod-limitrange.yml 
limitrange "limit-namespece" deleted
##运行一个容器
[root@master resources]# kubectl run myapp --image=kmckenzie/myapp:v1
kubectl run --generator=deployment/apps.v1 is DEPRECATED and will be removed in a future version. Use kubectl run --generator=run-pod/v1 or kubectl create instead.
deployment.apps/myapp created
[root@master resources]# kubectl get pod
myapp-769ffcfc6c-wq8p5        1/1     Running   0          4s


### 可以正常运行
[root@master resources]# kubectl apply -f pod-resourcequot.yaml 
resourcequota/pod-resourcequota created
##查看
[root@master resources]# kubectl describe resourcequota
Name:            pod-resourcequota
Namespace:       default
Resource         Used   Hard
--------         ----   ----
limits.cpu       1      2
limits.memory    300Mi  1Gi
requests.cpu     500m   1
requests.memory  100Mi  500Mi  ##可以看见使用了多少我们在多创建几个

4️⃣ 多创建几个

[root@master resources]# kubectl run myapp2 --image=kmckenzie/myapp:v1
kubectl run --generator=deployment/apps.v1 is DEPRECATED and will be removed in a future version. Use kubectl run --generator=run-pod/v1 or kubectl create instead.
deployment.apps/myapp2 created
[root@master resources]# kubectl run myapp3 --image=kmckenzie/myapp:v1
kubectl run --generator=deployment/apps.v1 is DEPRECATED and will be removed in a future version. Use kubectl run --generator=run-pod/v1 or kubectl create instead.
deployment.apps/myapp3 created
[root@master resources]# kubectl run myapp4 --image=kmckenzie/myapp:v1 
kubectl run --generator=deployment/apps.v1 is DEPRECATED and will be removed in a future version. Use kubectl run --generator=run-pod/v1 or kubectl create instead.
deployment.apps/myapp4 created


###会发现myapp3,4容器并没有起来

[root@master resources]# kubectl get pod
NAME                      READY   STATUS    RESTARTS   AGE
myapp-769ffcfc6c-lktz9    1/1     Running   0          87s
myapp2-6c4b4d6df9-4plkm   1/1     Running   0          17s

5️⃣ 再来查看一个 urcequota，发现只要达到其中一个限额他就不会继续创建成功

[root@master resources]#  kubectl describe resourcequota 
Name:            pod-resourcequota
Namespace:       default
Resource         Used   Hard
--------         ----   ----
limits.cpu       2      2  ##可以看见cpu已经达到限额了
limits.memory    600Mi  1Gi
requests.cpu     1      1
requests.memory  200Mi  500Mi

⭐️ 为 namespace创建 Pod限额

1️⃣ 先把之前的限额删掉

[root@master resources]# kubectl delete -f pod-limitrange.yml 
limitrange "limit-namespece" deleted
[root@master resources]# kubectl delete -f pod-resourcequot.yaml 
resourcequota "pod-resourcequota" deleted

2️⃣ 编写配置

[root@master resources]# cat namespec-pod.yaml 
apiVersion: v1
kind: ResourceQuota
metadata:
  name: resourcequota-pod
spec:
  hard:
    pods: "3"

3️⃣ 运行3个pod运行成功

[root@master resources]# kubectl run myapp3 --image=kmckenzie/myapp:v1
kubectl run --generator=deployment/apps.v1 is DEPRECATED and will be removed in a future version. Use kubectl run --generator=run-pod/v1 or kubectl create instead.
deployment.apps/myapp3 created
[root@master resources]# kubectl get pod
NAME                      READY   STATUS    RESTARTS   AGE
myapp-769ffcfc6c-lktz9    1/1     Running   0          11m
myapp2-6c4b4d6df9-4plkm   1/1     Running   0          10m
myapp3-664b99bb45-9fxsk   1/1     Running   0          3s

4️⃣ 运行4个 pod，发现无法运行启动

[root@master resources]# kubectl run myapp4 --image=kmckenzie/myapp:v1
kubectl run --generator=deployment/apps.v1 is DEPRECATED and will be removed in a future version. Use kubectl run --generator=run-pod/v1 or kubectl create instead.
deployment.apps/myapp4 create
[root@master resources]# kubectl get pod
NAME                      READY   STATUS    RESTARTS   AGE
myapp-769ffcfc6c-lktz9    1/1     Running   0          11m
myapp2-6c4b4d6df9-4plkm   1/1     Running   0          10m
myapp3-664b99bb45-9fxsk   1/1     Running   0          35s

site logo

Hi,Friend

© 2024 七点半的菜市场