前面学习了 ReplicaSet 控制器，了解到该控制器是用来维护集群中运行的 Pod 数量的，但是往往在实际操作的时候，反而不会去直接使用 RS，而是会使用更上层的控制器，比如今天要学习的主角 Deployment，Deployment 一个非常重要的功能就是实现了 Pod 的滚动更新，比如应用更新了，只需要更新容器镜像，然后修改 Deployment 里面的 Pod 模板镜像，那么 Deployment 就会用滚动更新（Rolling Update）的方式来升级现在的 Pod，这个能力是非常重要的，因为对于线上的服务需要做到不中断服务，所以滚动更新就成了必须的一个功能。而 Deployment 这个能力的实现，依赖的就是 ReplicaSet 这个资源对象，可以通俗的理解就是每个 Deployment 就对应集群中的一次部署。
Deployment
Deployment 资源对象的格式和 ReplicaSet 几乎一致，如下资源对象就是一个常见的 Deployment 资源类型：（nginx-deploy.yaml）
apiVersion: apps/v1kind: Deploymentmetadata:name: nginx-deploynamespace: defaultspec:replicas: 3 # 期望的 Pod 副本数量，默认值为1selector: # Label Selector，必须匹配 Pod 模板中的标签matchLabels:app: nginxtemplate: # Pod 模板metadata:labels:app: nginxspec:containers:- name: nginximage: nginxports:- containerPort: 80
这里只是将类型替换成了 Deployment，我们可以先来创建下这个资源对象：
$ kubectl apply -f nginx-deploy.yaml
deployment.apps/nginx-deploy created
$ kubectl get deployment
NAME READY UP-TO-DATE AVAILABLE AGE
nginx-deploy 3/3 3 3 58s
# 创建完成后，查看 Pod 状态：
$ kubectl get pods -l app=nginx
NAME READY STATUS RESTARTS AGE
nginx-deploy-85ff79dd56-7r76h 1/1 Running 0 41s
nginx-deploy-85ff79dd56-d5gjs 1/1 Running 0 41s
nginx-deploy-85ff79dd56-txc4h 1/1 Running 0 41s
到这里发现和之前的 RS 对象是否没有什么两样，都是根据spec.replicas来维持的副本数量，随意查看一个 Pod 的描述信息：
$ kubectl describe pod nginx-deploy-85ff79dd56-txc4h
Name: nginx-deploy-85ff79dd56-txc4h
Namespace: default
Priority: 0
PriorityClassName: <none>
Node: ydzs-node1/10.151.30.22
Start Time: Sat, 16 Nov 2019 16:01:25 +0800
Labels: app=nginx
pod-template-hash=85ff79dd56
Annotations: podpreset.admission.kubernetes.io/podpreset-time-preset: 2062768
Status: Running
IP: 10.244.1.166
Controlled By: ReplicaSet/nginx-deploy-85ff79dd56
......
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled <unknown> default-scheduler Successfully assigned default/nginx-deploy-85ff79dd56-txc4h to ydzs-node1
Normal Pulling 2m kubelet, ydzs-node1 Pulling image "nginx"
Normal Pulled 117s kubelet, ydzs-node1 Successfully pulled image "nginx"
Normal Created 117s kubelet, ydzs-node1 Created container nginx
Normal Started 116s kubelet, ydzs-node1 Started container nginx
仔细查看其中有这样一个信息Controlled By: ReplicaSet/nginx-deploy-85ff79dd56，什么意思？是不是表示当前这个 Pod 的控制器是一个 ReplicaSet 对象啊，我们不是创建的一个 Deployment 吗？为什么 Pod 会被 RS 所控制呢？那再去看下这个对应的 RS 对象的详细信息如何呢：
$ kubectl describe rs nginx-deploy-85ff79dd56
Name: nginx-deploy-85ff79dd56
Namespace: default
Selector: app=nginx,pod-template-hash=85ff79dd56
Labels: app=nginx
pod-template-hash=85ff79dd56
Annotations: deployment.kubernetes.io/desired-replicas: 3
deployment.kubernetes.io/max-replicas: 4
deployment.kubernetes.io/revision: 1
Controlled By: Deployment/nginx-deploy
Replicas: 3 current / 3 desired
Pods Status: 3 Running / 0 Waiting / 0 Succeeded / 0 Failed
......
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal SuccessfulCreate 4m52s replicaset-controller Created pod: nginx-deploy-85ff79dd56-7r76h
Normal SuccessfulCreate 4m52s replicaset-controller Created pod: nginx-deploy-85ff79dd56-d5gjs
Normal SuccessfulCreate 4m52s replicaset-controller Created pod: nginx-deploy-85ff79dd56-txc4h
其中有这样的一个信息：Controlled By: Deployment/nginx-deploy，明白了吧？意思就是 Pod 依赖的控制器 RS 实际上被Deployment 控制着呢，用下图来说明 Pod、ReplicaSet、Deployment 三者之间的关系：
通过上图我们可以很清楚的看到，定义了3个副本的 Deployment 与 ReplicaSet 和 Pod 的关系，就是一层一层进行控制的。ReplicaSet 作用和之前一样还是来保证 Pod 的个数始终保存指定的数量，所以 Deployment 中的容器 restartPolicy=Always 是唯一的就是这个原因，因为容器必须始终保证自己处于 Running 状态，ReplicaSet 才可以去明确调整 Pod 的个数。而 Deployment 是通过管理 ReplicaSet 的数量和属性来实现水平扩展/收缩以及滚动更新两个功能的。
水平伸缩
水平扩展/收缩的功能比较简单，因为ReplicaSet就可以实现，所以Deployment控制器只需要去修改他所控制的ReplicaSet的Pod副本数量就可以了。比如现在把Pod的副本数调整到4个，那么Deployment所对应ReplicaSet就会自动创建一个新的Pod出来，这就是水平扩展，使用一个新的命令kubectl scale来完成这个操作：
$ kubectl scale deployment nginx-deploy --replicas=4
deployment.apps/nginx-deployment scaled
扩容完成后可以查看当前的RS对象：
$ kubectl get rs
NAME DESIRED CURRENT READY AGE
nginx-deploy-85ff79dd56 4 4 3 40m
可以看到期望的Pod数量已经变成4了，只是Pod还没准备完成，所以READY状态数量还是3，同样查看RS的详细信息：
$ kubectl describe rs nginx-deploy-85ff79dd56
Name: nginx-deploy-85ff79dd56
Namespace: default
Selector: app=nginx,pod-template-hash=85ff79dd56
......
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal SuccessfulCreate 40m replicaset-controller Created pod: nginx-deploy-85ff79dd56-7r76h
Normal SuccessfulCreate 40m replicaset-controller Created pod: nginx-deploy-85ff79dd56-d5gjs
Normal SuccessfulCreate 40m replicaset-controller Created pod: nginx-deploy-85ff79dd56-txc4h
Normal SuccessfulCreate 17s replicaset-controller Created pod: nginx-deploy-85ff79dd56-tph9g
可以看到RS控制器增加了一个新的Pod，同样的DEployment资源对象的事件中也可以看到扩容完成的操作：
$ kubectl describe deploy nginx-deploy
Name: nginx-deploy
Namespace: default
......
OldReplicaSets: <none>
NewReplicaSet: nginx-deploy-85ff79dd56 (4/4 replicas created)
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal ScalingReplicaSet 43m deployment-controller Scaled up replica set nginx-deploy-85ff79dd56 to 3
Normal ScalingReplicaSet 3m16s deployment-controller Scaled up replica set nginx-deploy-85ff79dd56 to 4
滚动更新
如果只是水平扩展/收缩这两个功能，就完全没必要设计 Deployment 这个资源对象了，Deployment 最突出的一个功能是支持滚动更新，比如现在需要把应用容器更改为 nginx:1.7.9 版本，修改后的资源清单文件如下所示：
apiVersion: apps/v1
kind: Deployment
metadata:
name: nginx-deploy
namespace: default
spec:
replicas: 3
selector:
matchLabels:
app: nginx
minReadySeconds: 5
strategy:
type: RollingUpdate # 指定更新策略：RollingUpdate和Recreate
rollingUpdate:
maxSurge: 1
maxUnavailable: 1
template:
metadata:
labels:
app: nginx
spec:
containers:
- name: nginx
image: nginx:1.7.9
ports:
- containerPort: 80
和前面相比较，除了更改了镜像之外，我们还指定了更新策略：
minReadySeconds: 5
strategy:
type: RollingUpdate
rollingUpdate:
maxSurge: 1
maxUnavailable: 1
minReadySeconds：表示k8s在等待设置的时间后才进行升级，如果没有设置该值，k8s会假设该容器启动起来后就提供服务，在某些极端情况下可能会造成服务不正常运行，默认值为0。type=RollingUpdate：表示设置更新策略为滚动更新，可以设置为Recreate和RollingUpdate，Recreate表示全部重新创建，默认值为RollingUpdate。MaxSurge：表示升级过程中最多可以比原先设置多出的Pod数量，例如MaxSurge=1，replicas=5，就表示k8s会先启动一个新的Pod，然后才删掉一个旧Pod，整个升级过程最多会出现5+1个Pod。maxUnavailable：表示升级过程中最多有多少个Pod处于无法提供服务的状态，当maxSurge不为0时，该值也不能为0。
直接更新上面的Deployment资源对象：
$ kubectl apply -f nginx-deploy.yaml
更新后，执行下面的kubectl rollout status命令来查看此次更新的状态：
$ kubectl rollout status deployment/nginx-deploy
Waiting for deployment "nginx-deploy" rollout to finish: 2 out of 3 new replicas have been updated...
从上面的信息可以看出滚动更新已经有两个 Pod 已经更新完成了，在滚动更新过程中，还可以执行如下的命令来暂停更新：
$ kubectl rollout pause deployment/nginx-deploy
deployment.apps/nginx-deploy paused
这个时候滚动更新就暂停了，此时可以查看下 Deployment 的详细信息：
$ kubectl describe deploy nginx-deploy
Name: nginx-deploy
Namespace: default
CreationTimestamp: Sat, 16 Nov 2019 16:01:24 +0800
Labels: <none>
Annotations: deployment.kubernetes.io/revision: 2
kubectl.kubernetes.io/last-applied-configuration:
{"apiVersion":"apps/v1","kind":"Deployment","metadata":{"annotations":{},"name":"nginx-deploy","namespace":"default"},"spec":{"minReadySec...
Selector: app=nginx
Replicas: 3 desired | 2 updated | 4 total | 4 available | 0 unavailable
StrategyType: RollingUpdate
MinReadySeconds: 5
RollingUpdateStrategy: 1 max unavailable, 1 max surge
......
OldReplicaSets: nginx-deploy-85ff79dd56 (2/2 replicas created)
NewReplicaSet: nginx-deploy-5b7b9ccb95 (2/2 replicas created)
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal ScalingReplicaSet 26m deployment-controller Scaled up replica set nginx-deploy-85ff79dd56 to 4
Normal ScalingReplicaSet 3m44s deployment-controller Scaled down replica set nginx-deploy-85ff79dd56 to 3
Normal ScalingReplicaSet 3m44s deployment-controller Scaled up replica set nginx-deploy-5b7b9ccb95 to 1
Normal ScalingReplicaSet 3m44s deployment-controller Scaled down replica set nginx-deploy-85ff79dd56 to 2
Normal ScalingReplicaSet 3m44s deployment-controller Scaled up replica set nginx-deploy-5b7b9ccb95 to 2
这个过程就是滚动更新的过程，启动一个新的 Pod，杀掉一个旧的 Pod，然后再启动一个新的 Pod，这样滚动更新下去，直到全都变成新的 Pod，这个时候系统中应该存在 4 个 Pod，因为设置的策略maxSurge=1，所以在升级过程中是允许的，而且是两个新的 Pod，两个旧的 Pod：
$ kubectl get pods -l app=nginx
NAME READY STATUS RESTARTS AGE
nginx-deploy-5b7b9ccb95-k6pkh 1/1 Running 0 11m
nginx-deploy-5b7b9ccb95-l6lmx 1/1 Running 0 11m
nginx-deploy-85ff79dd56-7r76h 1/1 Running 0 75m
nginx-deploy-85ff79dd56-txc4h 1/1 Running 0 75m
查看 Deployment 的状态也可以看到当前的 Pod 状态：
$ kubectl get deployment
NAME READY UP-TO-DATE AVAILABLE AGE
nginx-deploy 4/3 2 4 75m
可以使用kubectl rollout resume来恢复滚动更新：
$ kubectl rollout resume deployment/nginx-deploy
deployment.apps/nginx-deploy resumed
$ kubectl rollout status deployment/nginx-deploy
Waiting for deployment "nginx-deploy" rollout to finish: 2 of 3 updated replicas are available...
deployment "nginx-deploy" successfully rolled out
看到上面的信息证明滚动更新已经成功了，同样可以查看下资源状态：
$ kubectl get pod -l app=nginx
NAME READY STATUS RESTARTS AGE
nginx-deploy-5b7b9ccb95-gmq7v 1/1 Running 0 115s
nginx-deploy-5b7b9ccb95-k6pkh 1/1 Running 0 15m
nginx-deploy-5b7b9ccb95-l6lmx 1/1 Running 0 15m
$ kubectl get deployment
NAME READY UP-TO-DATE AVAILABLE AGE
nginx-deploy 3/3 3 3 79m
查看 ReplicaSet 对象，可以发现会出现两个：
$ kubectl get rs -l app=nginx
NAME DESIRED CURRENT READY AGE
nginx-deploy-5b7b9ccb95 3 3 3 18m
nginx-deploy-85ff79dd56 0 0 0 81m
从上面可以看出滚动更新之前使用的 RS 资源对象的 Pod 副本数已经变成 0 了，而滚动更新后的 RS 资源对象变成了 3 个副本，导出之前的 RS 对象查看：
$ kubectl get rs nginx-deploy-85ff79dd56 -o yaml
apiVersion: apps/v1
kind: ReplicaSet
metadata:
annotations:
deployment.kubernetes.io/desired-replicas: "3"
deployment.kubernetes.io/max-replicas: "4"
deployment.kubernetes.io/revision: "1"
creationTimestamp: "2019-11-16T08:01:24Z"
generation: 5
labels:
app: nginx
pod-template-hash: 85ff79dd56
name: nginx-deploy-85ff79dd56
namespace: default
ownerReferences:
- apiVersion: apps/v1
blockOwnerDeletion: true
controller: true
kind: Deployment
name: nginx-deploy
uid: b0fc5614-ef58-496c-9111-740353bd90d4
resourceVersion: "2140545"
selfLink: /apis/apps/v1/namespaces/default/replicasets/nginx-deploy-85ff79dd56
uid: 8eca2998-3610-4f80-9c21-5482ba579892
spec:
replicas: 0
selector:
matchLabels:
app: nginx
pod-template-hash: 85ff79dd56
template:
metadata:
creationTimestamp: null
labels:
app: nginx
pod-template-hash: 85ff79dd56
spec:
containers:
- image: nginx
imagePullPolicy: Always
name: nginx
ports:
- containerPort: 80
protocol: TCP
resources: {}
terminationMessagePath: /dev/termination-log
terminationMessagePolicy: File
dnsPolicy: ClusterFirst
restartPolicy: Always
schedulerName: default-scheduler
securityContext: {}
terminationGracePeriodSeconds: 30
status:
observedGeneration: 5
replicas: 0
观察这个资源对象里面的描述信息除了副本数变成了replicas=0之外，和更新之前没有什么区别，有了这个 RS 的记录存在，就可以进行回滚，而且还可以回滚到前面的任意一个版本，通过命令 rollout history 来获取：
$ kubectl rollout history deployment nginx-deploy
deployment.apps/nginx-deploy
REVISION CHANGE-CAUSE
1 <none>
2 <none>
其实 rollout history 中记录的 revision 是和 ReplicaSets 一一对应。删除某个 ReplicaSet，对应的rollout history就会被删除，也就是说无法回滚到这个revison了，查看一个revison的详细信息：
$ kubectl rollout history deployment nginx-deploy --revision=1
deployment.apps/nginx-deploy with revision #1
Pod Template:
Labels: app=nginx
pod-template-hash=85ff79dd56
Containers:
nginx:
Image: nginx
Port: 80/TCP
Host Port: 0/TCP
Environment: <none>
Mounts: <none>
Volumes: <none>
假如现在要回退到当前版本的前一个版本，直接使用如下命令进行操作：
$ kubectl rollout undo deployment nginx-deploy
也可以回退到指定的revision版本：
$ kubectl rollout undo deployment nginx-deploy --to-revision=1
deployment "nginx-deploy" rolled back
回滚的过程中可以查看回滚状态：
$ kubectl rollout status deployment/nginx-deploy
Waiting for deployment "nginx-deploy" rollout to finish: 1 old replicas are pending termination...
Waiting for deployment "nginx-deploy" rollout to finish: 1 old replicas are pending termination...
Waiting for deployment "nginx-deploy" rollout to finish: 1 old replicas are pending termination...
Waiting for deployment "nginx-deploy" rollout to finish: 2 of 3 updated replicas are available...
Waiting for deployment "nginx-deploy" rollout to finish: 2 of 3 updated replicas are available...
deployment "nginx-deploy" successfully rolled out
这个时候查看对应的 RS 资源对象可以看到 Pod 副本已经回到之前的 RS 里面去了。
$ kubectl get rs -l app=nginx
NAME DESIRED CURRENT READY AGE
nginx-deploy-5b7b9ccb95 0 0 0 31m
nginx-deploy-85ff79dd56 3 3 3 95m
不过需要注意的是回滚的操作滚动的revision始终是递增的：
$ kubectl rollout history deployment nginx-deploy
deployment.apps/nginx-deploy
REVISION CHANGE-CAUSE
2 <none>
3 <none>
若有收获，就点个赞吧
0 人点赞
