第十一章 kubernetes 核心技术-Controller控制器
一、Controller控制器概述
1.含义
在集群上管理和运行容器的对象。
2.Controller和Pod之间的关系
Pod是通过Controller实现应用的运维比如伸缩,滚动升级等等
Pod和Controller之间通过label标签建立关系
二、Replication Controller
Replication Controller(RC)是 Kubernetes 系统中核心概念之一，当我们定义了一个 RC 并提交到 Kubernetes 集群中以后，Master 节点上的 Controller Manager 组件就得到通知， 定期检查系统中存活的 Pod,并确保目标 Pod 实例的数量刚好等于 RC 的预期值，如果有过多或过少的 Pod 运行，系统就会停掉或创建一些 Pod.此外我们也可以通过修改 RC 的副本数量，来实现 Pod 的动态缩放功能。
kubectl scale rc nginx --replicas=5
由于 Replication Controller 与 Kubernetes 代码中的模块 Replication Controller 同名，所以在 Kubernetes v1.2 时， 它就升级成了另外一个新的概念Replica Sets,官方解释为下一代的 RC，它与 RC 区别是:Replica Sets 支援基于集合的 Label selector,而 RC 只支持基于等式的 Label Selector。我们很少单独使用 Replica Set,它主要被 Deployment 这个更高层面的资源对象所使用，从而形成一整套 Pod 创建、删除、更新的编排机制。最好不要越过 RC 直接创建 Pod， 因为 Replication Controller 会通过 RC 管理 Pod 副本，实现自动创建、补足、替换、删除 Pod 副本，这样就能提高应用的容灾能力，减少由于节点崩溃等意外状况造成的损失。即使应用程序只有一个 Pod 副本，也强烈建议使用 RC 来 定义 Pod.
三、Replica Set
ReplicaSet 跟 ReplicationController 没有本质的不同，只是名字不一样，并且ReplicaSet 支持集合式的 selector（ReplicationController 仅支持等式）。
Kubernetes 官方强烈建议避免直接使用 ReplicaSet，而应该通过 Deployment 来创建 RS 和Pod。由于 ReplicaSet 是 ReplicationController 的代替物，因此用法基本相同，唯一的区别在于 ReplicaSet 支持集合式的 selector。
四、Deployment控制器
1.含义
Deployment 是 Kubenetes v1.2 引入的新概念，引入的目的是为Pod提供声明式更新。Deployment 内部使用了 Replica Set 来实现。Deployment 的定义与 Replica Set 的定义很类似，除了 API 声明与 Kind 类型有所区别.
在Deployment对象中描述所需的状态，然后Deployment控制器将实际状态以受控的速率更改为所需的状态。您可以定义部署以创建新的副本集，或删除现有部署并在新部署中采用其所有资源。一句话：Deployment主要功能是保证有足够的Pod正常对外提供服务。
2.Deployment应用场景
#1.部署无状态应用
#2.管理Pod和Replica Set部署,滚动升级等功能
#3.应用场景:web服务,微服务
3.Deployment实例
1）手写Deployment
[root@kubernetes-master-001 ~]# vim demo.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: nginx-deployment
labels:
app: nginx
spec:
replicas: 3
selector:
matchLabels:
app: nginx
template:
metadata:
labels:
app: nginx
spec:
containers:
- name: nginx
image: nginx
ports:
- containerPort: 80
2）导出Deployment模板
[root@kubernetes-master-001 ~]# kubectl create deployment web --image=nginx --dry-run=client -o yaml > web.yaml
[root@kubernetes-master-001 ~]# ll
total 20
-rw-r--r-- 1 root root 376 Nov 8 15:31 web.yaml
3）应用部署
将kubectl标志设置--record为true允许您将当前命令记录在正在创建或更新的资源的注释中。这对于将来的检查很有用。
[root@kubernetes-master-001 ~]# kubectl apply -f web.yaml --record
deployment.apps/web created
[root@kubernetes-master-001 ~]# kubectl get deployments.apps
NAME READY UP-TO-DATE AVAILABLE AGE
web 1/1 1 1 39s
[root@kubernetes-master-001 ~]# kubectl rollout history deployment web
deployment.apps/web
REVISION CHANGE-CAUSE
1
4）对外发布（暴露端口号）
[root@kubernetes-master-001 ~]# kubectl expose deployment web --port=80 --type=NodePort --target-port=80 --name=web1 -o yaml > web1.yaml
[root@kubernetes-master-001 ~]# kubectl apply -f web1.yaml
service/web1 configured
[root@kubernetes-master-001 ~]# kubectl get pod,svc
NAME READY STATUS RESTARTS AGE
pod/web-96d5df5c8-kgw6v 1/1 Running 0 3m58s
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/kubernetes ClusterIP 10.96.0.1 443/TCP 3h17m
service/web1 NodePort 10.98.97.60 80:31413/TCP 82s
5）验证访问
[root@kubernetes-master-001 ~]# curl 192.168.13.100:31413
Welcome to nginx!
Welcome to nginx!
If you see this page, the nginx web server is successfully installed and
working. Further configuration is required.
For online documentation and support please refer to
nginx.org.
Commercial support is available at
nginx.com.
Thank you for using nginx.
4.查看部署状态
[root@kubernetes-master-001 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-96d5df5c8-kgw6v 1/1 Running 0 23m
5.查看部署详情
[root@kubernetes-master-001 ~]# kubectl describe deployments.apps web
Name: web
Namespace: default
CreationTimestamp: Mon, 08 Nov 2021 17:22:44 +0800
Labels: app=web
Annotations: deployment.kubernetes.io/revision: 1
Selector: app=web
Replicas: 1 desired | 1 updated | 1 total | 1 available | 0 unavailable
StrategyType: RollingUpdate
MinReadySeconds: 0
RollingUpdateStrategy: 25% max unavailable, 25% max surge
Pod Template:
Labels: app=web
Containers:
nginx:
Image: nginx
Port:
Host Port:
Environment:
Mounts:
Volumes:
Conditions:
Type Status Reason
---- ------ ------
Available True MinimumReplicasAvailable
Progressing True NewReplicaSetAvailable
OldReplicaSets:
NewReplicaSet: web-96d5df5c8 (1/1 replicas created)
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal ScalingReplicaSet 25m deployment-controller Scaled up replica set web-96d5df5c8 to 1
五、Deployment应用升级回滚
一般对应用程序升级或者版本迭代时，会通过Deployment对Pod进行滚动更新。
1.修改配置文件
[root@kubernetes-master-001 ~]# vim web.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
creationTimestamp: null
labels:
app: web
name: web
spec:
replicas: 2
selector:
matchLabels:
app: web
strategy: {}
template:
metadata:
creationTimestamp: null
labels:
app: web
spec:
containers:
- image: nginx:1.14
name: nginx
resources: {}
status: {}
2.应用部署
#1.查看应用部署状态
[root@kubernetes-master-001 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-6c57bdf5f4-cwr8r 1/1 Running 0 4s
web-6c57bdf5f4-r57s9 1/1 Running 0 73s
#2.查看部署详情
[root@kubernetes-master-001 ~]# kubectl describe deployments.apps web
Name: web
Namespace: default
CreationTimestamp: Mon, 08 Nov 2021 17:55:40 +0800
Labels: app=web
Annotations: deployment.kubernetes.io/revision: 2
Selector: app=web
Replicas: 2 desired | 2 updated | 2 total | 2 available | 0 unavailable
StrategyType: RollingUpdate
MinReadySeconds: 0
RollingUpdateStrategy: 25% max unavailable, 25% max surge
Pod Template:
Labels: app=web
Containers:
nginx:
Image: nginx:1.14
Port:
Host Port:
Environment:
Mounts:
Volumes:
Conditions:
Type Status Reason
---- ------ ------
Available True MinimumReplicasAvailable
Progressing True NewReplicaSetAvailable
OldReplicaSets:
NewReplicaSet: web-5bb6fd4c98 (2/2 replicas created)
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal ScalingReplicaSet 5m51s deployment-controller Scaled up replica set web-6c57bdf5f4 to 1
Normal ScalingReplicaSet 4m42s deployment-controller Scaled up replica set web-6c57bdf5f4 to 2
Normal ScalingReplicaSet 3m9s deployment-controller Scaled up replica set web-5bb6fd4c98 to 1
Normal ScalingReplicaSet 2m41s deployment-controller Scaled down replica set web-6c57bdf5f4 to 1
Normal ScalingReplicaSet 2m41s deployment-controller Scaled up replica set web-5bb6fd4c98 to 2
Normal ScalingReplicaSet 2m38s deployment-controller Scaled down replica set web-6c57bdf5f4 to 0
3.应用升级
#1.修改nginx镜像1.14版本为1.18
[root@kubernetes-master-001 ~]# kubectl set image deployment web nginx=nginx:1.18
deployment.apps/web image updated
#2.直接修改部署应用镜像
[root@kubernetes-master-001 ~]# kubectl edit deployments.apps web
#2.查看升级状态
[root@kubernetes-master-001 ~]# kubectl rollout status deployment web
deployment "web" successfully rolled out
4.应用回滚
当新版本不稳定时，可以对其进行回滚操作，默认情况下，所有Deployment的rollout历史都保留在系统中，可以随时回滚。
#1.查看构建历史
[root@kubernetes-master-001 ~]# kubectl rollout history deployment web
deployment.apps/web
REVISION CHANGE-CAUSE
1
2
5.回滚到上一版本
#1.开始回滚
[root@kubernetes-master-001 ~]# kubectl rollout undo deployment web
deployment.apps/web rolled back
#2.查看构建状态
[root@kubernetes-master-001 ~]# kubectl rollout status deployment web
Waiting for deployment "web" rollout to finish: 1 old replicas are pending termination...
Waiting for deployment "web" rollout to finish: 1 old replicas are pending termination...
deployment "web" successfully rolled out
#2.查看新的构建历史， 发现历史2被删除，新增了历史3
[root@kubernetes-master-001 ~]# kubectl rollout history deployment web
deployment.apps/web
REVISION CHANGE-CAUSE
1
3
6.回滚到指定版本
#1.查看当前历史版本
[root@kubernetes-master-001 ~]# kubectl rollout history deployment web
deployment.apps/web
REVISION CHANGE-CAUSE
3
4
=
#2.回滚至指定版本
[root@kubernetes-master-001 ~]# kubectl rollout undo deployment web --to-revision=4
deployment.apps/web skipped rollback (current template already matches revision 4)
#3.查看回滚状态
[root@kubernetes-master-001 ~]# kubectl describe deployments.apps web
Name: web
Namespace: default
CreationTimestamp: Mon, 08 Nov 2021 17:55:40 +0800
Labels: app=web
Annotations: deployment.kubernetes.io/revision: 4
Selector: app=web
Replicas: 2 desired | 2 updated | 2 total | 2 available | 0 unavailable
StrategyType: RollingUpdate
MinReadySeconds: 0
RollingUpdateStrategy: 25% max unavailable, 25% max surge
Pod Template:
Labels: app=web
Containers:
nginx:
Image: nginx:1.14
Port:
Host Port:
Environment:
Mounts:
Volumes:
Conditions:
Type Status Reason
---- ------ ------
Available True MinimumReplicasAvailable
Progressing True NewReplicaSetAvailable
OldReplicaSets:
NewReplicaSet: web-5bb6fd4c98 (2/2 replicas created)
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal ScalingReplicaSet 26m (x2 over 34m) deployment-controller Scaled up replica set web-6c57bdf5f4 to 1
Normal ScalingReplicaSet 26m (x2 over 33m) deployment-controller Scaled up replica set web-6c57bdf5f4 to 2
Normal ScalingReplicaSet 26m deployment-controller Scaled down replica set web-5bb6fd4c98 to 1
Normal ScalingReplicaSet 26m deployment-controller Scaled down replica set web-5bb6fd4c98 to 0
Normal ScalingReplicaSet 14m (x2 over 31m) deployment-controller Scaled up replica set web-5bb6fd4c98 to 1
Normal ScalingReplicaSet 14m (x2 over 31m) deployment-controller Scaled down replica set web-6c57bdf5f4 to 1
Normal ScalingReplicaSet 14m (x2 over 31m) deployment-controller Scaled up replica set web-5bb6fd4c98 to 2
Normal ScalingReplicaSet 14m (x2 over 31m) deployment-controller Scaled down replica set web-6c57bdf5f4 to 0
六、Deployment弹性伸缩和扩容
当业务的用户越来越多，目前的后端服务已经无法满足业务要求当前的业务要求，传统的解决办法是将其横向增加服务器，从而满足我们的业务要求。K8S中也是支持横向扩容的方法的。
1.扩容
#1.第一种方式：scale
[root@kubernetes-master-001 ~]# kubectl scale deployment web --replicas=5
deployment.apps/web scaled
[root@kubernetes-master-001 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-5bb6fd4c98-7cgjs 1/1 Running 0 18s
web-5bb6fd4c98-8mn4k 1/1 Running 0 18s
web-5bb6fd4c98-9ssxq 1/1 Running 0 16h
web-5bb6fd4c98-wrd7d 1/1 Running 0 18s
web-5bb6fd4c98-zgvhz 1/1 Running 0 16h
[root@kubernetes-master-001 ~]# kubectl get deployments.apps web
NAME READY UP-TO-DATE AVAILABLE AGE
web 5/5 5 5 16h
#2.第二种方式：patch
[root@kubernetes-master-001 ~]# kubectl patch deployments.apps web -p '{"spec": {"replicas": 5}}'
deployment.apps/web patched
[root@kubernetes-master-001 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-5bb6fd4c98-9ssxq 1/1 Running 0 16h
web-5bb6fd4c98-bqcmt 1/1 Running 0 7s
web-5bb6fd4c98-pvpc8 1/1 Running 0 7s
web-5bb6fd4c98-tqnzx 1/1 Running 0 7s
web-5bb6fd4c98-zgvhz 1/1 Running 0 16h
[root@kubernetes-master-001 ~]# kubectl get deployments.apps web
NAME READY UP-TO-DATE AVAILABLE AGE
web 5/5 5 5 16h
2.暂停部署
#1.暂停部署
[root@kubernetes-master-01 test]# kubectl rollout pause deployment nginx-deployment
error: deployments.apps "nginx-deployment" is already paused
#2.设置镜像升级
[root@kubernetes-master-001 ~]# kubectl set image deployment web nginx=nginx:1.20
deployment.apps/web image updated
#3.查看部署状态，发现没有开始部署
[root@kubernetes-master-001 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-5bb6fd4c98-9ssxq 1/1 Running 0 16h
web-5bb6fd4c98-bqcmt 1/1 Running 0 6m19s
web-5bb6fd4c98-pvpc8 1/1 Running 0 6m19s
web-5bb6fd4c98-tqnzx 1/1 Running 0 6m19s
web-5bb6fd4c98-zgvhz 1/1 Running 0 16h
#4.取消暂停，开始部署
[root@kubernetes-master-001 ~]# kubectl rollout resume deployment web
deployment.apps/web resumed
#5.查看状态，发现开始部署
[root@kubernetes-master-001 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-5bb6fd4c98-9ssxq 1/1 Running 0 16h
web-5bb6fd4c98-bqcmt 1/1 Running 0 6m55s
web-5bb6fd4c98-pvpc8 1/1 Running 0 6m55s
web-5bb6fd4c98-tqnzx 0/1 Terminating 0 6m55s
web-5bb6fd4c98-zgvhz 1/1 Running 0 16h
web-7f5844f89c-94w65 0/1 ContainerCreating 0 2s
web-7f5844f89c-ggqz5 0/1 ContainerCreating 0 2s
web-7f5844f89c-tw8rd 0/1 ContainerCreating 0 2s
#6.查看应用镜像
[root@kubernetes-master-001 ~]# kubectl describe deployments.apps web |grep Image
Image: nginx:1.20
七、Horizontal Pod Autoscaler
在生产环境中，总会有一些意想不到的事情发生，比如公司网站流量突然升高，此时之前创建的Pod已不足以撑住所有的访问，而运维人员也不可能24小时守着业务服务，这时就可以通过配置HPA，实现负载过高的情况下自动扩容Pod副本数以分摊高并发的流量，当流量恢复正常后，HPA会自动缩减Pod的数量。HPA是根据CPU的使用率、内存使用率自动扩展Pod数量的，所以要使用HPA就必须定义Requests参数。
Horizontal Pod Autoscal(Pod 横向扩容 简称 HPA)与 RC、Deployment 一样，也属于一种 Kubernetes 资源对象。通过追踪分析 RC 控制的所有目标 Pod 的负载变化情况，来确定是否需要针对性地调整目标 Pod 的副本数，这是 HPA 的 实现原理。
Kubernetes 对 Pod 扩容与缩容提供了手动和自动两种模式，手动模式通过 kubectl scale 命令对一个 Deployment/RC 进行 Pod 副本数量的设置。自动模式则需要用户根据某个性能指标或者自定义业务指标，并指定 Pod 副本数量的范围，系统将自动在这个范围内根据性能指标的变化进行调整。
1.手动扩容和缩容
[root@kubernetes-master-001 ~]# kubectl scale deployment web --replicas=1
deployment.apps/web scaled
[root@kubernetes-master-001 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-7f5844f89c-94w65 0/1 Terminating 0 15m
web-7f5844f89c-ggqz5 0/1 Terminating 0 15m
web-7f5844f89c-lc6r5 0/1 Terminating 0 14m
web-7f5844f89c-tw8rd 1/1 Running 0 15m
web-7f5844f89c-wmrsq 0/1 Terminating 0 14m
[root@kubernetes-master-001 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-7f5844f89c-tw8rd 1/1 Running 0 15m
2.自动扩容和缩容
HPA 控制器基本 Master 的 kube-controller-manager 服务启动参数 --horizontal-pod- autoscaler-sync-period 定义的时长(默认值为 30s),周期性地监测 Pod 的 CPU 使用率， 并在满足条件时对 RC 或 Deployment 中的 Pod 副 本数量进行调整，以符合用户定义的平均Pod CPU 使用率。
#1.创建HPA
kind: Deployment
apiVersion: apps/v1
metadata:
name: nginx
namespace: default
labels:
app: nginx
spec:
selector:
matchLabels:
app: nginx
template:
metadata:
labels:
app: nginx
spec:
containers:
- name: nginx
image: nginx
ports:
- containerPort: 80
resources:
limits:
cpu: 10m
memory: 50Mi
requests:
cpu: 10m
memory: 50Mi
---
kind: Service
apiVersion: v1
metadata:
name: svc
namespace: default
spec:
selector:
app: nginx
ports:
- port: 80
targetPort: 80
---
kind: HorizontalPodAutoscaler
apiVersion: autoscaling/v2beta1
metadata:
name: docs
namespace: default
spec:
# HPA的最小pod数量和最大pod数量
maxReplicas: 10
minReplicas: 1
# HPA的伸缩对象描述，HPA会动态修改该对象的pod数量
scaleTargetRef:
kind: Deployment
name: nginx
apiVersion: apps/v1
# 监控的指标数组，支持多种类型的指标共存
metrics:
- type: Resource
# 核心指标，包含cpu和内存两种（被弹性伸缩的pod对象中容器的requests和limits中定义的指标。）
resource:
name: cpu
# CPU 阈值
# 计算公式：所有目标pod的metric的使用率（百分比）的平均值，
# 例如limit.cpu=1000m，实际使用500m，则utilization=50%
# 例如deployment.replica=3, limit.cpu=1000m，则pod1实际使用cpu=500m, pod2=300m, pod=600m
## 则averageUtilization=(500/1000+300/1000+600/1000)/3 = (500 + 300 + 600)/(3*1000)）
targetAverageUtilization: 5
[root@kubernetes-master-001 ~]# kubectl run test --rm -it --image=busybox --command -- sh
If you don't see a command prompt, try pressing enter.
/ # while true; do wget -O- -q 10.96.62.55; done
[root@kubernetes-master-001 ~]# kubectl get pods -o wide -w
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
hook-demo1 1/1 Running 0 20h 10.242.112.3 kubernetes-node-02
nginx-7bf65b57b6-6sfp5 1/1 Running 0 5m6s 10.240.120.3 kubernetes-node-03
tcp-pod 1/1 Running 0 36m 10.242.0.3 kubernetes-node-05
test 1/1 Running 0 2m3s 10.240.120.4 kubernetes-node-03
nginx-7bf65b57b6-2qwn9 0/1 Pending 0 0s
nginx-7bf65b57b6-2qwn9 0/1 Pending 0 0s kubernetes-node-01
nginx-7bf65b57b6-2qwn9 0/1 ContainerCreating 0 0s kubernetes-node-01
[root@kubernetes-master-001 ~]# kubectl top pods
NAME CPU(cores) MEMORY(bytes)
nginx-7bf65b57b6-2qwn9 10m 4Mi
nginx-7bf65b57b6-6sfp5 10m 4Mi
nginx-7bf65b57b6-h8gp2 0m 0Mi
nginx-7bf65b57b6-mb2rz 0m 0Mi