k8s学习--kubernetes服务自动伸缩之水平收缩(pod副本收缩)VPA策略应用案例

爱吃龙利鱼

于 2024-06-10 10:12:38 发布

阅读量1.1k

点赞数 32

分类专栏： k8s 文章标签： kubernetes 学习容器云原生

本文链接：https://blog.csdn.net/lwxvgdv/article/details/139564746

版权

k8s 专栏收录该内容

24 篇文章 0 订阅

订阅专栏

文章目录

前言
应用
- 环境

前言

有任何疑问或不懂的地方均可评论或私信，欢迎交流
关于VPA的详细解释
链接: VPA的详细解释

策略
在VPA中，updateMode 是一个重要的配置选项，它决定了VPA如何应用其提供的资源建议。根据不同的设置，VPA可以采取不同的策略来更新Pod的资源配置：

Off：
VPA不会应用任何资源推荐，只是收集和显示数据。

Auto
概述:
自动调整策略在无需重启 Pod 的情况下动态调整其资源请求。

特性:
动态调整： Pod 的资源请求在运行时逐步增加。

优点：
无需重启 Pod 即可调整资源，最小化中断。
提供更及时的资源调整，适合短周期和不稳定的负载。

缺点：
由于容器运行时资源的硬限制，可能无法完全满足新资源请求，导致资源不足。

Recreate
概述:
当资源需求变化时，使用重建策略重新启动 Pod 以调整资源请求。

特性:
重启调整：Pod 会被杀死并重新创建，以适应新的资源请求。

优点：
确保 Pod 能获取到新的资源请求，避免运行时的资源限制问题。
适合稳定负载且可以容忍短暂停机的应用。

缺点：
重启 Pod 会引起短暂的服务中断。
不适合要求高可用性且不能容忍重启的工作负载。

Initial
概述:
初始调整策略仅在 Pod 第一次创建时设置资源请求，不会对运行中的 Pod 进行调整。
特性:
静态调整：在 Pod 创建时基于历史数据设定初始资源请求。

优点：
避免了运行时调整带来的复杂性。
适合长期运行的负载，初始设置资源合理即可。

缺点：
无法调整已经运行的 Pod 的资源，若需求变化则需手动干预。

应用

环境

虚拟机

Ip	主机名	cpu	内存	硬盘
192.168.10.11	master01	2cpu双核	4G	100G
192.168.10.12	worker01	2cpu双核	4G	100G
192.168.10.13	worker02	2cpu双核	4G	100G

版本 centos7.9
已部署k8s-1.27

1.VPA应用案例 updateMode: “Off”

（1）创建应用实例

VPA不会应用任何资源推荐，只是收集和显示数据。

vim 03-nginx.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: nginx
  name: nginx
  namespace: default
spec:
  replicas: 2
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx:latest 
        imagePullPolicy: IfNotPresent
        resources:
          requests:
            cpu: 100m
            memory: 250Mi
---
apiVersion: v1
kind: Service
metadata:
  name: nginx
  namespace: default
spec:
  type: NodePort
  ports:
  - port: 80
    targetPort: 80
  selector:
    app: nginx

kubectl apply -f 03-nginx.yaml
kubectl get pods

在这里插入图片描述

kubectl get svc

在这里插入图片描述

（2）创建vpa

使用updateMode: "Off"模式，这种模式仅获取资源推荐，不更新Pod

 vim nginx-vpa.yaml

apiVersion: autoscaling.k8s.io/v1
kind: VerticalPodAutoscaler
metadata:
  name: nginx-vpa
  namespace: default
spec:
  targetRef:
    apiVersion: "apps/v1"
    kind: Deployment
    name: nginx
  updatePolicy:
    updateMode: "Off"
  resourcePolicy:
    containerPolicies:
    - containerName: "nginx"
      minAllowed:
        cpu: "250m"
        memory: "100Mi"
      maxAllowed:
        cpu: "2000m"
        memory: "2048Mi"

kubectl apply -f nginx-vpa.yaml
 kubectl get vpa

稍等片刻
在这里插入图片描述

kubectl describe vpa nginx-vpa

Name:         nginx-vpa
Namespace:    default
Labels:       <none>
Annotations:  <none>
API Version:  autoscaling.k8s.io/v1
Kind:         VerticalPodAutoscaler
Metadata:
  Creation Timestamp:  2024-06-09T11:10:34Z
  Generation:          1
  Resource Version:    19449
  UID:                 be19f937-fb0c-4c33-a559-4e5aa52043b8
Spec:
  Resource Policy:
    Container Policies:
      Container Name:  nginx
      Max Allowed:
        Cpu:     2000m
        Memory:  2048Mi
      Min Allowed:
        Cpu:     250m
        Memory:  100Mi
  Target Ref:
    API Version:  apps/v1
    Kind:         Deployment
    Name:         nginx
  Update Policy:
    Update Mode:  Off
Status:
  Conditions:
    Last Transition Time:  2024-06-09T11:11:17Z
    Status:                True
    Type:                  RecommendationProvided
  Recommendation:
    Container Recommendations:
      Container Name:  nginx
      Lower Bound:
        Cpu:     250m
        Memory:  262144k
      Target:
        Cpu:     250m
        Memory:  262144k
      Uncapped Target:
        Cpu:     25m
        Memory:  262144k
      Upper Bound:
        Cpu:     1142m
        Memory:  1194357142
Events:          <none>

解释如下：
Recommendation:：包含对Pod资源需求的推荐值。
Container Recommendations:：针对特定容器的推荐值。
Container Name:：容器名称。
Lower Bound:：推荐的下限资源量。
Target:：推荐的最优资源量。
Uncapped Target:：如果没有上限约束，则为目标资源量。
Upper Bound:：推荐的上限资源量。

kubectl get svc

在这里插入图片描述

yum -y install httpd-tools
ab -c 1000 -n 100000000 http://192.168.10.11:30478/

打开一个新终端

kubectl describe vpa nginx-vpa

Name:         nginx-vpa
Namespace:    default
Labels:       <none>
Annotations:  <none>
API Version:  autoscaling.k8s.io/v1
Kind:         VerticalPodAutoscaler
Metadata:
  Creation Timestamp:  2024-06-09T11:10:34Z
  Generation:          1
  Resource Version:    20221
  UID:                 be19f937-fb0c-4c33-a559-4e5aa52043b8
Spec:
  Resource Policy:
    Container Policies:
      Container Name:  nginx
      Max Allowed:
        Cpu:     2000m
        Memory:  2048Mi
      Min Allowed:
        Cpu:     250m
        Memory:  100Mi
  Target Ref:
    API Version:  apps/v1
    Kind:         Deployment
    Name:         nginx
  Update Policy:
    Update Mode:  Off
Status:
  Conditions:
    Last Transition Time:  2024-06-09T11:11:17Z
    Status:                True
    Type:                  RecommendationProvided
  Recommendation:
    Container Recommendations:
      Container Name:  nginx
      Lower Bound:
        Cpu:     250m
        Memory:  262144k
      Target:
        Cpu:     250m
        Memory:  262144k
      Uncapped Target:
        Cpu:     25m
        Memory:  262144k
      Upper Bound:
        Cpu:     802m
        Memory:  838810574
Events:          <none>

由于使用updateMode: “Off”，所以没有更新pod

kubectl get pods

在这里插入图片描述

2.VPA应用案例 updateMode: “Auto”

此模式当目前运行的pod的资源达不到VPA的推荐值，就会执行pod驱逐，重新部署新的足够资源的服务

（1）创建应用

vim 05-nginx.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: nginx
  name: nginx
  namespace: default
spec:
  replicas: 2
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx:latest
        imagePullPolicy: IfNotPresent
        resources:
          requests:
            cpu: 100m
            memory: 50Mi
---
apiVersion: v1
kind: Service
metadata:
  name: nginx
  namespace: default
spec:
  type: NodePort
  ports:
  - port: 80
    targetPort: 80
  selector:
    app: nginx

kubectl apply -f 05-nginx.yaml
 kubectl get pods

 kubectl apply -f 05-nginx.yaml
 kubectl get pods

在这里插入图片描述

（2）创建vpa

vim nginx-vpa-auto.yaml

apiVersion: autoscaling.k8s.io/v1
kind: VerticalPodAutoscaler
metadata:
  name: nginx-vpa-auto
  namespace: default
spec:
  targetRef:
    apiVersion: "apps/v1"
    kind: Deployment
    name: nginx
  updatePolicy:
    updateMode: "Auto"
  resourcePolicy:
    containerPolicies:
    - containerName: "nginx"
      minAllowed:
        cpu: "250m"
        memory: "100Mi"
      maxAllowed:
        cpu: "2000m"
        memory: "2048Mi"

 kubectl apply -f nginx-vpa-auto.yaml
kubectl get vpa

在这里插入图片描述

（3）执行压测

kubectl get svc

在这里插入图片描述

ab -c 1000 -n 1000000000 http://192.168.10.11:31115/

打开另一个终端

 kubectl describe vpa nginx-vpa-auto

Name:         nginx-vpa-auto
Namespace:    default
Labels:       <none>
Annotations:  <none>
API Version:  autoscaling.k8s.io/v1
Kind:         VerticalPodAutoscaler
Metadata:
  Creation Timestamp:  2024-06-09T11:27:07Z
  Generation:          1
  Resource Version:    21845
  UID:                 453b03aa-fba5-4bcd-aa66-f85e43a57521
Spec:
  Resource Policy:
    Container Policies:
      Container Name:  nginx
      Max Allowed:
        Cpu:     2000m
        Memory:  2048Mi
      Min Allowed:
        Cpu:     250m
        Memory:  100Mi
  Target Ref:
    API Version:  apps/v1
    Kind:         Deployment
    Name:         nginx
  Update Policy:
    Update Mode:  Auto
Status:
  Conditions:
    Last Transition Time:  2024-06-09T11:27:17Z
    Status:                True
    Type:                  RecommendationProvided
  Recommendation:
    Container Recommendations:
      Container Name:  nginx
      Lower Bound:
        Cpu:     250m
        Memory:  262144k
      Target:
        Cpu:     250m
        Memory:  262144k
      Uncapped Target:
        Cpu:     25m
        Memory:  262144k
      Upper Bound:
        Cpu:     521m
        Memory:  545693548
Events:          <none>

 kubectl get event

LAST SEEN   TYPE      REASON                   OBJECT                        MESSAGE
38m         Normal    Scheduled                pod/nginx-59d7c8bd89-q27gc    Successfully assigned default/nginx-59d7c8bd89-q27gc to worker02
38m         Normal    Pulling                  pod/nginx-59d7c8bd89-q27gc    Pulling image "nginx:latest"
38m         Normal    Pulled                   pod/nginx-59d7c8bd89-q27gc    Successfully pulled image "nginx:latest" in 18.118818504s (18.118826419s including waiting)
38m         Normal    Created                  pod/nginx-59d7c8bd89-q27gc    Created container nginx
38m         Normal    Started                  pod/nginx-59d7c8bd89-q27gc    Started container nginx
11m         Normal    Killing                  pod/nginx-59d7c8bd89-q27gc    Stopping container nginx
38m         Normal    Scheduled                pod/nginx-59d7c8bd89-qb4p4    Successfully assigned default/nginx-59d7c8bd89-qb4p4 to worker01
38m         Normal    Pulling                  pod/nginx-59d7c8bd89-qb4p4    Pulling image "nginx:latest"
38m         Normal    Pulled                   pod/nginx-59d7c8bd89-qb4p4    Successfully pulled image "nginx:latest" in 17.842596583s (17.842603536s including waiting)
38m         Normal    Created                  pod/nginx-59d7c8bd89-qb4p4    Created container nginx
38m         Normal    Started                  pod/nginx-59d7c8bd89-qb4p4    Started container nginx
11m         Normal    Killing                  pod/nginx-59d7c8bd89-qb4p4    Stopping container nginx
38m         Normal    SuccessfulCreate         replicaset/nginx-59d7c8bd89   Created pod: nginx-59d7c8bd89-q27gc
38m         Normal    SuccessfulCreate         replicaset/nginx-59d7c8bd89   Created pod: nginx-59d7c8bd89-qb4p4
11m         Normal    Scheduled                pod/nginx-6f78fbb759-hcbc6    Successfully assigned default/nginx-6f78fbb759-hcbc6 to worker02
11m         Normal    Pulled                   pod/nginx-6f78fbb759-hcbc6    Container image "nginx:latest" already present on machine
11m         Normal    Created                  pod/nginx-6f78fbb759-hcbc6    Created container nginx
11m         Normal    Started                  pod/nginx-6f78fbb759-hcbc6    Started container nginx
10m         Normal    Killing                  pod/nginx-6f78fbb759-hcbc6    Stopping container nginx
9m51s       Normal    Scheduled                pod/nginx-6f78fbb759-jhf7h    Successfully assigned default/nginx-6f78fbb759-jhf7h to worker02
9m51s       Normal    Pulled                   pod/nginx-6f78fbb759-jhf7h    Container image "nginx:latest" already present on machine
9m51s       Normal    Created                  pod/nginx-6f78fbb759-jhf7h    Created container nginx
9m51s       Normal    Started                  pod/nginx-6f78fbb759-jhf7h    Started container nginx
6m31s       Normal    Killing                  pod/nginx-6f78fbb759-jhf7h    Stopping container nginx
6m31s       Normal    EvictedByVPA             pod/nginx-6f78fbb759-jhf7h    Pod was evicted by VPA Updater to apply resource recommendation.
6m30s       Normal    Scheduled                pod/nginx-6f78fbb759-mcppm    Successfully assigned default/nginx-6f78fbb759-mcppm to worker02
6m30s       Normal    Pulled                   pod/nginx-6f78fbb759-mcppm    Container image "nginx:latest" already present on machine
6m30s       Normal    Created                  pod/nginx-6f78fbb759-mcppm    Created container nginx
6m30s       Normal    Started                  pod/nginx-6f78fbb759-mcppm    Started container nginx
9m51s       Normal    Scheduled                pod/nginx-6f78fbb759-s4xzm    Successfully assigned default/nginx-6f78fbb759-s4xzm to worker01
9m51s       Normal    Pulled                   pod/nginx-6f78fbb759-s4xzm    Container image "nginx:latest" already present on machine
9m51s       Normal    Created                  pod/nginx-6f78fbb759-s4xzm    Created container nginx
9m51s       Normal    Started                  pod/nginx-6f78fbb759-s4xzm    Started container nginx
7m31s       Normal    EvictedByVPA             pod/nginx-6f78fbb759-s4xzm    Pod was evicted by VPA Updater to apply resource recommendation.
7m31s       Normal    Killing                  pod/nginx-6f78fbb759-s4xzm    Stopping container nginx
7m30s       Normal    Scheduled                pod/nginx-6f78fbb759-x986h    Successfully assigned default/nginx-6f78fbb759-x986h to worker01
7m30s       Normal    Pulled                   pod/nginx-6f78fbb759-x986h    Container image "nginx:latest" already present on machine
7m30s       Normal    Created                  pod/nginx-6f78fbb759-x986h    Created container nginx
7m30s       Normal    Started                  pod/nginx-6f78fbb759-x986h    Started container nginx
11m         Normal    Scheduled                pod/nginx-6f78fbb759-z4sxb    Successfully assigned default/nginx-6f78fbb759-z4sxb to worker01
11m         Normal    Pulled                   pod/nginx-6f78fbb759-z4sxb    Container image "nginx:latest" already present on machine
11m         Normal    Created                  pod/nginx-6f78fbb759-z4sxb    Created container nginx
11m         Normal    Started                  pod/nginx-6f78fbb759-z4sxb    Started container nginx
10m         Normal    Killing                  pod/nginx-6f78fbb759-z4sxb    Stopping container nginx
11m         Normal    SuccessfulCreate         replicaset/nginx-6f78fbb759   Created pod: nginx-6f78fbb759-hcbc6
11m         Normal    SuccessfulCreate         replicaset/nginx-6f78fbb759   Created pod: nginx-6f78fbb759-z4sxb
9m52s       Normal    SuccessfulCreate         replicaset/nginx-6f78fbb759   Created pod: nginx-6f78fbb759-jhf7h
9m52s       Normal    SuccessfulCreate         replicaset/nginx-6f78fbb759   Created pod: nginx-6f78fbb759-s4xzm
7m31s       Normal    SuccessfulCreate         replicaset/nginx-6f78fbb759   Created pod: nginx-6f78fbb759-x986h
6m31s       Normal    SuccessfulCreate         replicaset/nginx-6f78fbb759   Created pod: nginx-6f78fbb759-mcppm
38m         Normal    ScalingReplicaSet        deployment/nginx              Scaled up replica set nginx-59d7c8bd89 to 2
11m         Normal    ScalingReplicaSet        deployment/nginx              Scaled up replica set nginx-6f78fbb759 to 2
10m         Warning   FailedToUpdateEndpoint   endpoints/nginx               Failed to update endpoint default/nginx: Operation cannot be fulfilled on endpoints "nginx": the object has been modified; please apply your changes to the latest version and try again
9m52s       Normal    ScalingReplicaSet        deployment/nginx              Scaled up replica set nginx-6f78fbb759 to 2

从输出信息可以了解到，vpa执行了EvictedByVPA，自动停掉了nginx，然后使用 VPA推荐的资源启动了新的nginx，我们查看下nginx的pod可以得到确认

kubectl describe pods nginx-6f78fbb759-mcppm

Name:             nginx-6f78fbb759-mcppm
Namespace:        default
Priority:         0
Service Account:  default
Node:             worker02/192.168.10.13
Start Time:       Sun, 09 Jun 2024 19:29:09 +0800
Labels:           app=nginx
                  pod-template-hash=6f78fbb759
Annotations:      cni.projectcalico.org/containerID: 145e1f8509982da00e61eb180bb9f3b52e527bc94d8657237d340fd9bf01475a
                  cni.projectcalico.org/podIP: 10.244.30.72/32
                  cni.projectcalico.org/podIPs: 10.244.30.72/32
                  vpaObservedContainers: nginx
                  vpaUpdates: Pod resources updated by nginx-vpa-auto: container 0: cpu request, memory request
Status:           Running
IP:               10.244.30.72
IPs:
  IP:           10.244.30.72
Controlled By:  ReplicaSet/nginx-6f78fbb759
Containers:
  nginx:
    Container ID:   docker://97f865221289a48d042dfd039b2eb33876c2d08a89c2f7863fd7e42b5f7f8018
    Image:          nginx:latest
    Image ID:       docker-pullable://nginx@sha256:0f04e4f646a3f14bf31d8bc8d885b6c951fdcf42589d06845f64d18aec6a3c4d
    Port:           <none>
    Host Port:      <none>
    State:          Running
      Started:      Sun, 09 Jun 2024 19:29:10 +0800
    Ready:          True
    Restart Count:  0
    Requests:
      cpu:        250m
      memory:     262144k
    Environment:  <none>
    Mounts:
      /var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-gfzhm (ro)
Conditions:
  Type              Status
  Initialized       True 
  Ready             True 
  ContainersReady   True 
  PodScheduled      True 
Volumes:
  kube-api-access-gfzhm:
    Type:                    Projected (a volume that contains injected data from multiple sources)
    TokenExpirationSeconds:  3607
    ConfigMapName:           kube-root-ca.crt
    ConfigMapOptional:       <nil>
    DownwardAPI:             true
QoS Class:                   Burstable
Node-Selectors:              <none>
Tolerations:                 node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
                             node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:
  Type    Reason     Age    From               Message
  ----    ------     ----   ----               -------
  Normal  Scheduled  8m56s  default-scheduler  Successfully assigned default/nginx-6f78fbb759-mcppm to worker02
  Normal  Pulled     8m56s  kubelet            Container image "nginx:latest" already present on machine
  Normal  Created    8m56s  kubelet            Created container nginx
  Normal  Started    8m56s  kubelet            Started container nginx

随着服务的负载的变化，VPA的推荐值也会不断变化。当目前运行的pod的资源达不到VPA的推荐值，就会执行pod驱逐，重新部署新的足够资源的服务。

爱吃龙利鱼

关注

32
点赞
踩
8

收藏

觉得还不错? 一键收藏
0
评论
k8s学习--kubernetes服务自动伸缩之水平收缩(pod副本收缩)VPA策略应用案例

有任何疑问或不懂的地方均可评论或私信，欢迎交流关于VPA的详细解释链接:VPA的详细解释策略在VPA中，updateMode 是一个重要的配置选项，它决定了VPA如何应用其提供的资源建议。VPA不会应用任何资源推荐，只是收集和显示数据。Auto自动调整策略在无需重启 Pod 的情况下动态调整其资源请求。动态调整： Pod 的资源请求在运行时逐步增加。无需重启 Pod 即可调整资源，最小化中断。提供更及时的资源调整，适合短周期和不稳定的负载。
复制链接

扫一扫