1、RC(Replication Controller)副本控制器，Replication Controller的作用。
应用托管在kubernetes之后，kubernetes需要保证应用能够持续运行，这是RC的工作内容，它会确保任何时间kubernetes中都有指定数量的Pod在运行。在此基础上，RC还提供了一些更高级的特性，比如滚动升级，升级回滚等等。
通俗的理解就是，当将应用通过K8s运行起来之后，我们的k8s是需要保证容器一直处于持续运行，保证它的高可用，那么我们的RC就可以确保容器的高可用，RC的工作原理就是，RC是会一直监控我们的K8S容器，也就是说POD资源它的运行状态，一旦发现这个Pod资源有异常了，那么我们的RC就会控制k8s在其他的Node节点上启动一个新的Pod，以此来保证这个业务的高可用运行。RC除了保证Pod高可用之外，还提供了更高级的特性，比如滚动升级，升级回滚等等。
2、首先，查看你的k8s各个节点状态是否正常运行，然后创建一个rc的目录，用于存放RC(Replication Controller)的yaml配置文件。
1 [root@k8s-master ~]# kubectl get nods 2 the server doesn‘t have a resource type "nods" 3 [root@k8s-master ~]# kubectl get node 4 NAME STATUS AGE 5 k8s-master Ready 6d 6 k8s-node2 Ready 6d 7 k8s-node3 Ready 6d 8 [root@k8s-master ~]# kubectl get nodes 9 NAME STATUS AGE 10 k8s-master Ready 6d 11 k8s-node2 Ready 6d 12 k8s-node3 Ready 6d 13 [root@k8s-master ~]# clear 14 [root@k8s-master ~]# kubectl get componentstatus 15 NAME STATUS MESSAGE ERROR 16 scheduler Healthy ok 17 controller-manager Healthy ok 18 etcd-0 Healthy {"health":"true"} 19 [root@k8s-master ~]# clear 20 [root@k8s-master ~]# cd k8s/ 21 [root@k8s-master k8s]# ls 22 pod 23 [root@k8s-master k8s]# mkdir rc 24 [root@k8s-master k8s]# cd rc/ 25 [root@k8s-master rc]# ls 26 [root@k8s-master rc]# vim nginx_rc_yaml 27 [root@k8s-master rc]# kubectl create -f nginx_rc_yaml 28 replicationcontroller "myweb" created 29 [root@k8s-master rc]# kubectl get rc 30 NAME DESIRED CURRENT READY AGE 31 myweb 2 2 1 13s 32 [root@k8s-master rc]#
创建nginx_rc_yaml配置文件，配置内容，如下所示。
1 # 声明api的版本。 2 apiVersion: v1 3 # kind代表资源的类型，资源是ReplicationController。 4 kind: ReplicationController 5 # 资源叫什么名字，是在其属性metadata里面的。 6 metadata: 7 # 第一个属性name的值是myweb，即ReplicationController的名字就叫做myweb。 8 name: myweb 9 # spec是详细，详细里面定义了一个容器。 10 spec: 11 # 声明副本数量是2，代表了RC会启动两个相同的Pod。 12 replicas: 2 13 # 选择器。 14 selector: 15 app: myweb 16 # Pod的启动模板，和Pod的yaml配置信息基本差不多的，几乎一样，但是这里没有名称，是因为两个Pod名称不能完全一样的。 17 # 没有指定名称，RC会随机生成一个名称。 18 template: 19 # 资源叫什么名字，是在其属性metadata里面的。但是这里让RC随机生成指定数量的名称。 20 metadata: 21 # 给Pod贴上了一个标签，标签是app: web，标签是有一定的作用的。 22 labels: 23 app: myweb 24 # spec是详细，详细里面定义了一个容器。 25 spec: 26 # 定义一个容器，可以声明多个容器的。 27 containers: 28 # 容器的名称叫做myweb 29 - name: myweb 30 # 使用了什么镜像，可以使用官方公有的，也可以使用私有的。 31 image: 192.168.110.133:5000/nginx:1.13 32 # ports定义容器的端口。 33 ports: 34 # 容器的端口是80，如果容器有多个端口，可以在后面接着写一行即可。 35 - containerPort: 80
配置，如下所示：
如果如何控制yaml的格式，可以使用notepad++的yaml语言格式，或者在线yaml解析，或者idea的yaml配置文件，idea的yaml配置文件也推荐使用哦。
创建好RC(Replication Controller)之后，可以进行检查。可以看到RC创建了两个Pod，可以查看一下Pod的数量和状态。
1 [root@k8s-master rc]# kubectl get rc 2 NAME DESIRED CURRENT READY AGE 3 myweb 2 2 1 7m 4 [root@k8s-master rc]# kubectl get pods 5 NAME READY STATUS RESTARTS AGE 6 myweb-0hqc5 0/1 ImagePullBackOff 0 8m 7 myweb-2np4k 1/1 Running 0 8m 8 nginx 1/1 Running 1 3d 9 test1 0/1 ImagePullBackOff 0 2d 10 test2 2/2 Running 1 2d 11 test4 1/1 Running 0 2d 12 [root@k8s-master rc]#
很明显，我这里创建的两个Pod，有一个启动失败了。此时，我想将失败的Pod删除掉，但是我删除了一个，RC又帮助你启动了一个，嗯，真的是高可用啊，然后我将RC删除掉，这两个Pod就随着被删除掉了。
1 [root@k8s-master ~]# kubectl get rc 2 NAME DESIRED CURRENT READY AGE 3 myweb 2 2 1 17m 4 [root@k8s-master ~]# kubectl get pod -o wide 5 NAME READY STATUS RESTARTS AGE IP NODE 6 myweb-8cp7h 0/1 ImagePullBackOff 0 5m 172.16.85.3 k8s-master 7 myweb-qcgjl 1/1 Running 1 14m 172.16.5.2 k8s-node2 8 nginx 1/1 Running 2 3d 172.16.38.3 k8s-node3 9 test1 0/1 ImagePullBackOff 0 2d 172.16.85.2 k8s-master 10 test2 2/2 Running 3 2d 172.16.38.2 k8s-node3 11 test4 1/1 Running 1 2d 172.16.5.3 k8s-node2 12 [root@k8s-master ~]# kubectl delete rc myweb 13 replicationcontroller "myweb" deleted 14 [root@k8s-master ~]# kubectl get rc 15 No resources found. 16 [root@k8s-master ~]#
这里我将没有用的测试Pod都删除掉，因为我笔记本只有8g内存，可能内存不够用了。搞了一个小时，不是内存的问题，是之前部署k8s的时候，测试nginx的时候将nginx拼错了，尴尬。
1 [root@k8s-node2 ~]# docker images 2 REPOSITORY TAG IMAGE ID CREATED SIZE 3 docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB 4 192.168.110.133:5000/nginx 1.13 ae513a47849c 2 years ago 109 MB 5 docker.io/nginx 1.13 ae513a47849c 2 years ago 109 MB 6 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB 7 docker.io/tianyebj/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB 8 [root@k8s-node2 ~]#
而我的RC的yaml的配置文件，如下所示：
1 Version: v1 2 kind: ReplicationController 3 metadata: 4 name: myweb 5 spec: 6 replicas: 2 7 selector: 8 app: myweb 9 template: 10 metadata: 11 labels: 12 app: myweb 13 spec: 14 containers: 15 - name: myweb 16 image: 192.168.110.133:5000/nginx:1.13 17 # imagePullPolicy: IfNotPresent 18 ports: 19 - containerPort: 80
但是主节点的docker镜像是192.168.110.133:5000/ngnix，造成了每次创建RC，在主节点的Pod都无法启动，尴尬，还排查了这么久。真打脸。
1 [root@k8s-master ~]# docker images 2 REPOSITORY TAG IMAGE ID CREATED SIZE 3 docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB 4 docker.io/registry latest 708bc6af7e5e 4 months ago 25.8 MB 5 192.168.110.133:5000/ngnix 1.13 ae513a47849c 2 years ago 109 MB 6 registry.access.redhat.com/rhel7/pod-infrastructure latest 99965fb98423 2 years ago 209 MB 7 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB 8 [root@k8s-master ~]#
报错信息，也贴一下吧，方便自己以后使用，如下所示：
1 [root@k8s-master ~]# kubectl describe pod myweb-qwgsf 2 Name: myweb-qwgsf 3 Namespace: default 4 Node: k8s-master/192.168.110.133 5 Start Time: Thu, 11 Jun 2020 17:21:45 +0800 6 Labels: app=myweb 7 Status: Pending 8 IP: 172.16.85.2 9 Controllers: ReplicationController/myweb 10 Containers: 11 myweb: 12 Container ID: 13 Image: 192.168.110.133:5000/nginx:1.13 14 Image ID: 15 Port: 80/TCP 16 State: Waiting 17 Reason: ErrImagePull 18 Ready: False 19 Restart Count: 0 20 Volume Mounts: <none> 21 Environment Variables: <none> 22 Conditions: 23 Type Status 24 Initialized True 25 Ready False 26 PodScheduled True 27 No volumes. 28 QoS Class: BestEffort 29 Tolerations: <none> 30 Events: 31 FirstSeen LastSeen Count From SubObjectPath Type Reason Message 32 --------- -------- ----- ---- ------------- -------- ------ ------- 33 12m 12m 1 {kubelet k8s-master} Warning MissingClusterDNS kubelet does not have ClusterDNS IP configured and cannot create Pod using "ClusterFirst" policy. Falling back to DNSDefault policy. 34 12m 12m 1 {default-scheduler } Normal Scheduled Successfully assigned myweb-qwgsf to k8s-master 35 12m 6m 6 {kubelet k8s-master} spec.containers{myweb} Normal Pulling pulling image "192.168.110.133:5000/nginx:1.13" 36 12m 6m 6 {kubelet k8s-master} spec.containers{myweb} Warning Failed Failed to pull image "192.168.110.133:5000/nginx:1.13": Error while pulling image: Get http://192.168.110.133:5000/v1/repositories/nginx/images: dial tcp 192.168.110.133:5000: connect: connection refused 37 12m 6m 6 {kubelet k8s-master} Warning FailedSync Error syncing pod, skipping: failed to "StartContainer" for "myweb" with ErrImagePull: "Error while pulling image: Get http://192.168.110.133:5000/v1/repositories/nginx/images: dial tcp 192.168.110.133:5000: connect: connection refused" 38 39 12m 4m 31 {kubelet k8s-master} spec.containers{myweb} Normal BackOff Back-off pulling image "192.168.110.133:5000/nginx:1.13" 40 12m 4m 31 {kubelet k8s-master} Warning FailedSync Error syncing pod, skipping: failed to "StartContainer" for "myweb" with ImagePullBackOff: "Back-off pulling image \"192.168.110.133:5000/nginx:1.13\"" 41 42 32s 32s 1 {kubelet k8s-master} Warning MissingClusterDNS kubelet does not have ClusterDNS IP configured and cannot create Pod using "ClusterFirst" policy. Falling back to DNSDefault policy. 43 31s 20s 2 {kubelet k8s-master} spec.containers{myweb} Normal Pulling pulling image "192.168.110.133:5000/nginx:1.13" 44 31s 20s 2 {kubelet k8s-master} spec.containers{myweb} Warning Failed Failed to pull image "192.168.110.133:5000/nginx:1.13": Error while pulling image: Get http://192.168.110.133:5000/v1/repositories/nginx/images: dial tcp 192.168.110.133:5000: connect: connection refused 45 31s 20s 2 {kubelet k8s-master} Warning FailedSync Error syncing pod, skipping: failed to "StartContainer" for "myweb" with ErrImagePull: "Error while pulling image: Get http://192.168.110.133:5000/v1/repositories/nginx/images: dial tcp 192.168.110.133:5000: connect: connection refused" 46 47 30s 8s 2 {kubelet k8s-master} spec.containers{myweb} Normal BackOff Back-off pulling image "192.168.110.133:5000/nginx:1.13" 48 30s 8s 2 {kubelet k8s-master} Warning FailedSync Error syncing pod, skipping: failed to "StartContainer" for "myweb" with ImagePullBackOff: "Back-off pulling image \"192.168.110.133:5000/nginx:1.13\""
此处，将主节点的Docker镜像删除掉。
1 [root@k8s-master ~]# docker images 2 REPOSITORY TAG IMAGE ID CREATED SIZE 3 docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB 4 docker.io/registry latest 708bc6af7e5e 4 months ago 25.8 MB 5 192.168.110.133:5000/ngnix 1.13 ae513a47849c 2 years ago 109 MB 6 registry.access.redhat.com/rhel7/pod-infrastructure latest 99965fb98423 2 years ago 209 MB 7 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB 8 [root@k8s-master ~]# docker ps 9 CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES 10 [root@k8s-master ~]# docker ps -a 11 CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES 12 a27987d97039 registry "/entrypoint.sh /e..." 5 days ago Exited (2) 3 days ago registry 13 ee95778bd5d9 busybox "sh" 6 days ago Exited (127) 6 days ago friendly_payne 14 6d459781a3e5 busybox "sh" 6 days ago Exited (137) 5 days ago gracious_nightingale 15 [root@k8s-master ~]# docker rmi -f ae513a47849c 16 Untagged: 192.168.110.133:5000/ngnix:1.13 17 Untagged: 192.168.110.133:5000/ngnix@sha256:e4f0474a75c510f40b37b6b7dc2516241ffa8bde5a442bde3d372c9519c84d90 18 Deleted: sha256:ae513a47849c895a155ddfb868d6ba247f60240ec8495482eca74c4a2c13a881 19 Deleted: sha256:160a8bd939a9421818f499ba4fbfaca3dd5c86ad7a6b97b6889149fd39bd91dd 20 Deleted: sha256:f246685cc80c2faa655ba1ec9f0a35d44e52b6f83863dc16f46c5bca149bfefc 21 Deleted: sha256:d626a8ad97a1f9c1f2c4db3814751ada64f60aed927764a3f994fcd88363b659 22 [root@k8s-master ~]# docker images 23 REPOSITORY TAG IMAGE ID CREATED SIZE 24 docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB 25 docker.io/registry latest 708bc6af7e5e 4 months ago 25.8 MB 26 registry.access.redhat.com/rhel7/pod-infrastructure latest 99965fb98423 2 years ago 209 MB 27 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB 28 [root@k8s-master ~]# docker images 29 REPOSITORY TAG IMAGE ID CREATED SIZE 30 docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB 31 docker.io/registry latest 708bc6af7e5e 4 months ago 25.8 MB 32 registry.access.redhat.com/rhel7/pod-infrastructure latest 99965fb98423 2 years ago 209 MB 33 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB 34 [root@k8s-master ~]#
此时，将三台节点重启了，或者重启服务，我这里直接重启了三台机器，其所有服务全部重启。
1 [root@k8s-master ~]# free -h 2 total used free shared buff/cache available 3 Mem: 2.2G 880M 453M 12M 953M 1.1G 4 Swap: 2.0G 0B 2.0G 5 [root@k8s-master ~]# kubectl get node 6 NAME STATUS AGE 7 k8s-master Ready 6d 8 k8s-node2 Ready 6d 9 k8s-node3 Ready 6d 10 [root@k8s-master ~]# kubectl get componentstatus 11 NAME STATUS MESSAGE ERROR 12 scheduler Healthy ok 13 controller-manager Healthy ok 14 etcd-0 Healthy {"health":"true"} 15 [root@k8s-master ~]# kubectl get rc 16 NAME DESIRED CURRENT READY AGE 17 myweb 2 2 2 16m 18 [root@k8s-master ~]# docker images 19 REPOSITORY TAG IMAGE ID CREATED SIZE 20 docker.io/nginx latest 2622e6cca7eb 41 hours ago 132 MB 21 docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB 22 docker.io/registry latest 708bc6af7e5e 4 months ago 25.8 MB 23 192.168.110.133:5000/nginx 1.13 ae513a47849c 2 years ago 109 MB 24 docker.io/nginx 1.13 ae513a47849c 2 years ago 109 MB 25 registry.access.redhat.com/rhel7/pod-infrastructure latest 99965fb98423 2 years ago 209 MB 26 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB 27 [root@k8s-master ~]# kubectl get node -o wide 28 NAME STATUS AGE EXTERNAL-IP 29 k8s-master Ready 6d <none> 30 k8s-node2 Ready 6d <none> 31 k8s-node3 Ready 6d <none> 32 [root@k8s-master ~]# kubectl get pod -o wide 33 NAME READY STATUS RESTARTS AGE IP NODE 34 myweb-2h8b1 1/1 Running 1 17m 172.16.85.2 k8s-master 35 myweb-lfkmp 1/1 Running 1 17m 172.16.5.2 k8s-node2 36 test4 1/1 Running 1 13m 172.16.38.2 k8s-node3 37 [root@k8s-master ~]#
RC(Replication Controller)会始终保持Pod的数量为2，可以自己删除一个Pod，k8s的RC(Replication Controller)会里面帮助你启动一个新的Pod，RC(Replication Controller)会时刻监控Pod的状态，少了就启动，多了就进行删除，数量和配置文件yaml的数量一致。
3、RC(Replication Controller)如何与Pod进行关联呢？
答：使用到的标签Label（标签选择器）。在nginx_rc.yaml配置文件中，定义了RC的选择器是通过标签app:myweb来选择的，每一个Pod在运行的时候会自动加上一个标签叫做app:myweb，这样的话，RC会自动根据标签来选择我们的Pod。
可以通过命令kubectl describe pod myweb-2h8b1来查看标签。
1 [root@k8s-master rc]# kubectl describe pod myweb-2h8b1 2 Name: myweb-2h8b1 3 Namespace: default 4 Node: k8s-master/192.168.110.133 5 Start Time: Thu, 11 Jun 2020 17:51:06 +0800 6 Labels: app=myweb 7 Status: Running 8 IP: 172.16.85.2 9 Controllers: ReplicationController/myweb 10 Containers: 11 myweb: 12 Container ID: docker://27a9e6dfb65be540bb50c98d820a5b773c0ed01d09d2350baf6027cdf9e22257 13 Image: 192.168.110.133:5000/nginx:1.13 14 Image ID: docker-pullable://docker.io/nginx@sha256:b1d09e9718890e6ebbbd2bc319ef1611559e30ce1b6f56b2e3b479d9da51dc35 15 Port: 80/TCP 16 State: Running 17 Started: Thu, 11 Jun 2020 18:06:34 +0800 18 Last State: Terminated 19 Reason: Completed 20 Exit Code: 0 21 Started: Thu, 11 Jun 2020 18:01:41 +0800 22 Finished: Thu, 11 Jun 2020 18:05:42 +0800 23 Ready: True 24 Restart Count: 1 25 Volume Mounts: <none> 26 Environment Variables: <none> 27 Conditions: 28 Type Status 29 Initialized True 30 Ready True 31 PodScheduled True 32 No volumes. 33 QoS Class: BestEffort 34 Tolerations: <none> 35 No events. 36 [root@k8s-master rc]#
通过查看RC的标签。
1 [root@k8s-master rc]# kubectl get rc -o wide 2 NAME DESIRED CURRENT READY AGE CONTAINER(S) IMAGE(S) SELECTOR 3 myweb 2 2 2 3h myweb 192.168.110.133:5000/nginx:1.13 app=myweb 4 [root@k8s-master rc]#
RC(Replication Controller)是通过标签（标签选择器）来选择Pod，通过标签来决定这个Pod是归我来管理的。
通过修改创建好的Pod可以测试，如果多于指定数量的Pod数量，就会被删除掉，注意，删除掉的Pod是最年轻的那个Pod。kubectl edit pod test4命令可以修改创建好的Pod。
1 [root@k8s-master rc]# kubectl get all 2 NAME DESIRED CURRENT READY AGE 3 rc/myweb 2 2 2 3h 4 5 NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE 6 svc/kubernetes 10.254.0.1 <none> 443/TCP 6d 7 8 NAME READY STATUS RESTARTS AGE 9 po/myweb-2h8b1 1/1 Running 1 3h 10 po/myweb-lfkmp 1/1 Running 1 3h 11 po/test4 1/1 Running 1 3h 12 [root@k8s-master rc]# kubectl edit pod test4 13 pod "test4" edited 14 [root@k8s-master rc]# kubectl get all 15 NAME DESIRED CURRENT READY AGE 16 rc/myweb 2 2 2 3h 17 18 NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE 19 svc/kubernetes 10.254.0.1 <none> 443/TCP 6d 20 21 NAME READY STATUS RESTARTS AGE 22 po/myweb-lfkmp 1/1 Running 1 3h 23 po/test4 1/1 Running 1 3h 24 [root@k8s-master rc]#
4、RC(Replication Controller)的滚动升级。
答：滚动升级是一种平滑过渡的升级方式，通过逐步替换的策略，保证整体系统的稳定，在初始升级的时候就可以及时发现，调整问题，以保证问题影响度不好扩大。Kubernetes中滚动升级的命令如下所示：
首先，这里将配置文件nginx_rc.yaml进行拷贝，然后修改配置文件nginx_rc2.yaml，将myweb替换为myweb2。
1 [root@k8s-master rc]# cp nginx_rc.yaml nginx_rc2.yaml 2 [root@k8s-master rc]# ls 3 nginx_rc2.yaml nginx_rc.yaml 4 [root@k8s-master rc]# vim nginx_rc2.yaml 5 [root@k8s-master rc]#
具体配置，如下所示：
将myweb替换为myweb2，替换过后，然后将镜像版本修改为latest版本，如下所示：
将Nginx的latest版本镜像拉取下来docker pull docker.io/nginx:latest。然后将镜像上传到私有仓库里面，方便下载。
这里需要注意的是，我之前在配置Docker镜像加速的时候，在三台机器的vim /etc/sysconfig/docker。
1 [root@k8s-node3 ~]# vim /etc/sysconfig/docker
我在这个配置文件里面加的镜像加速和配置私有仓库地址。貌似并不是很好使的。
下面，在三台机器的上面，进行如下配置，将Docker镜像加速和私有仓库配置到下面这里。
1 [root@k8s-node3 ~]# docker pull docker.io/nginx:1.15 2 Trying to pull repository docker.io/library/nginx ... 3 Get https://registry-1.docker.io/v2/: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)
修改配置文件，在三台机器上面都配置如下所示配置，[root@k8s-node3 ~]# vim /etc/docker/daemon.json
1 { "insecure-registries":["192.168.110.133:5000"] ,"registry-mirrors":["https://docker.mirrors.ustc.edu.cn"]}
然后重启三台机器的Docker服务systemctl restart docker。如果实在下载不下来，需要自己从网上找个https://hub.docker.com/
1 [root@k8s-master ~]# docker pull docker.io/nginx:1.15 2 Trying to pull repository docker.io/library/nginx ... 3 sha256:23b4dcdf0d34d4a129755fc6f52e1c6e23bb34ea011b315d87e193033bcd1b68: Pulling from docker.io/library/nginx 4 743f2d6c1f65: Pull complete 5 6bfc4ec4420a: Pull complete 6 688a776db95f: Pull complete 7 Digest: sha256:23b4dcdf0d34d4a129755fc6f52e1c6e23bb34ea011b315d87e193033bcd1b68 8 Status: Downloaded newer image for docker.io/nginx:1.15 9 [root@k8s-master ~]#
然后将ngnix1.15上传到私有仓库里面。可以使用docker images命令查看是否已经上传到私有仓库。
1 [root@k8s-master rc]# docker images 2 REPOSITORY TAG IMAGE ID CREATED SIZE 3 docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB 4 docker.io/registry latest 708bc6af7e5e 4 months ago 25.8 MB 5 docker.io/nginx 1.15 53f3fd8007f7 13 months ago 109 MB 6 192.168.110.133:5000/nginx 1.13 ae513a47849c 2 years ago 109 MB 7 docker.io/nginx 1.13 ae513a47849c 2 years ago 109 MB 8 registry.access.redhat.com/rhel7/pod-infrastructure latest 99965fb98423 2 years ago 209 MB 9 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB 10 [root@k8s-master rc]# docker tag docker.io/nginx:1.15 192.168.110.133:5000/nginx:1.15 11 [root@k8s-master rc]# docker push 192.168.110.133:5000/nginx:1.15 12 The push refers to a repository [192.168.110.133:5000/nginx] 13 Put http://192.168.110.133:5000/v1/repositories/nginx/: dial tcp 192.168.110.133:5000: connect: connection refused 14 [root@k8s-master rc]# docker images 15 REPOSITORY TAG IMAGE ID CREATED SIZE 16 docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB 17 docker.io/registry latest 708bc6af7e5e 4 months ago 25.8 MB 18 docker.io/nginx 1.15 53f3fd8007f7 13 months ago 109 MB 19 192.168.110.133:5000/nginx 1.15 53f3fd8007f7 13 months ago 109 MB 20 docker.io/nginx 1.13 ae513a47849c 2 years ago 109 MB 21 192.168.110.133:5000/nginx 1.13 ae513a47849c 2 years ago 109 MB 22 registry.access.redhat.com/rhel7/pod-infrastructure latest 99965fb98423 2 years ago 209 MB 23 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB 24 [root@k8s-master rc]#
滚动升级是一种平滑过渡的升级方式，通过逐步替换的策略，保证整体系统的稳定，在初始升级的时候就可以及时发现，调整问题，以保证问题影响度不好扩大。Kubernetes中滚动升级的命令如下所示：
这里需要注意的是，想要看看Docker私有仓库是否有你想要的镜像，可以使用如下所示查看：
1 [root@k8s-node3 docker]# docker images 2 REPOSITORY TAG IMAGE ID CREATED SIZE 3 docker.io/busybox latest 1c35c4412082 9 days ago 1.22 MB 4 192.168.110.133:5000/nginx 1.15 53f3fd8007f7 13 months ago 109 MB 5 docker.io/nginx 1.15 53f3fd8007f7 13 months ago 109 MB 6 192.168.110.133:5000/nginx 1.13 ae513a47849c 2 years ago 109 MB 7 docker.io/tianyebj/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB 8 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB 9 [root@k8s-node3 docker]#
还有关于使用镜像加速和私有仓库的配置，这里需要说明的是，我的Docker的版本是1.13.1。
1 [root@k8s-node3 ~]# docker version 2 Client: 3 Version: 1.13.1 4 API version: 1.26 5 Package version: docker-1.13.1-161.git64e9980.el7_8.x86_64 6 Go version: go1.10.3 7 Git commit: 64e9980/1.13.1 8 Built: Tue Apr 28 14:43:01 2020 9 OS/Arch: linux/amd64 10 11 Server: 12 Version: 1.13.1 13 API version: 1.26 (minimum version 1.12) 14 Package version: docker-1.13.1-161.git64e9980.el7_8.x86_64 15 Go version: go1.10.3 16 Git commit: 64e9980/1.13.1 17 Built: Tue Apr 28 14:43:01 2020 18 OS/Arch: linux/amd64 19 Experimental: false 20 [root@k8s-node3 ~]#
关于镜像加速和私有仓库的配置，我此时由于不需要从网上下载所需的软件，这里将三台机器的/etc/docker/daemon.json配置文件全部后面加上了_bak，这里不需要使用它们了。
1 [root@k8s-node3 ~]# cat /etc/docker/daemon.json_bak 2 { "insecure-registries":["192.168.110.133:5000"]} 3 [root@k8s-node3 ~]# cd /etc/docker/ 4 [root@k8s-node3 docker]# ls 5 certs.d daemon.json_20200612 daemon.json_bak key.json seccomp.json 6 [root@k8s-node3 docker]#
关于镜像加速和私有仓库的配置，如果配置不好，真的难为si你的。所以这里说了很多次。此时，三台机器的镜像加速和私有中心在这个里面还进行了配置。特此说明。
1 [root@k8s-master docker]# cat /etc/sysconfig/docker 2 # /etc/sysconfig/docker 3 4 # Modify these options if you want to change the way the docker daemon runs 5 # OPTIONS=‘--selinux-enabled --log-driver=journald --signature-verification=false‘ 6 # 信任私有仓库，镜像加速 7 OPTIONS=‘--selinux-enabled --log-driver=journald --signature-verification=false 8 --registry-mirror=https://registry.docker-cn.com --insecure-registry=192.168.110.133:5000‘ 9 10 11 12 if [ -z "${DOCKER_CERT_PATH}" ]; then 13 DOCKER_CERT_PATH=/etc/docker 14 fi 15 16 # Do not add registries in this file anymore. Use /etc/containers/registries.conf 17 # instead. For more information reference the registries.conf(5) man page. 18 19 # Location used for temporary files, such as those created by 20 # docker load and build operations. Default is /var/lib/docker/tmp 21 # Can be overriden by setting the following environment variable. 22 # DOCKER_TMPDIR=/var/tmp 23 24 # Controls the /etc/cron.daily/docker-logrotate cron job status. 25 # To disable, uncomment the line below. 26 # LOGROTATE=false 27 28 # docker-latest daemon can be used by starting the docker-latest unitfile. 29 # To use docker-latest client, uncomment below lines 30 #DOCKERBINARY=/usr/bin/docker-latest 31 #DOCKERDBINARY=/usr/bin/dockerd-latest 32 #DOCKER_CONTAINERD_BINARY=/usr/bin/docker-containerd-latest 33 #DOCKER_CONTAINERD_SHIM_BINARY=/usr/bin/docker-containerd-shim-latest 34 [root@k8s-master docker]#
饶了一大圈，此时进行滚动升级，滚动升级是一种平滑过渡的升级方式，通过逐步替换的策略，保证整体系统的稳定，在初始升级的时候就可以及时发现，调整问题，以保证问题影响度不好扩大。Kubernetes中滚动升级的命令如下所示：
1 [root@k8s-master rc]# kubectl rolling-update myweb -f nginx_rc2.yaml --update-period=10s 2 Created myweb2 3 Scaling up myweb2 from 0 to 2, scaling down myweb from 2 to 0 (keep 2 pods available, don‘t exceed 3 pods) 4 Scaling myweb2 up to 1 5 Scaling myweb down to 1 6 Scaling myweb2 up to 2 7 Scaling myweb down to 0 8 Update succeeded. Deleting myweb 9 replicationcontroller "myweb" rolling updated to "myweb2" 10 [root@k8s-master rc]#
创建一个myweb2的RC，将myweb2的RC的Pod数量由0调整为2，把myweb的RC的Pod数量由2调整为0。当myweb2存活了30秒以上就会删掉一个myweb，nginx也是myweb的容器，删除也是比较慢的。
升级开始后，首先依据提供的定义文件创建V2版本的RC，然后每隔10s（--update-period=10s）逐步的增加V2版本的Pod副本数，逐步减少V1版本Pod的副本数。升级完成之后，删除V1版本的RC，保留V2版本的RC，以及实现滚动升级。
1 [root@k8s-master ~]# kubectl get pod -o wide 2 NAME READY STATUS RESTARTS AGE IP NODE 3 myweb2-f5400 1/1 Running 0 15s 172.16.38.3 k8s-node3 4 myweb2-mg9sk 1/1 Running 0 26s 172.16.85.2 k8s-master 5 [root@k8s-master ~]#
升级之后还可以进行回滚，如下所示：
1 [root@k8s-master rc]# kubectl rolling-update myweb2 -f nginx_rc.yaml --update-period=10s 2 Created myweb 3 Scaling up myweb from 0 to 2, scaling down myweb2 from 2 to 0 (keep 2 pods available, don‘t exceed 3 pods) 4 Scaling myweb up to 1 5 Scaling myweb2 down to 1 6 Scaling myweb up to 2 7 Scaling myweb2 down to 0 8 Update succeeded. Deleting myweb2 9 replicationcontroller "myweb2" rolling updated to "myweb" 10 [root@k8s-master rc]#
升级过程中，发生了错误中途退出时候，可以选择继续升级。Kubernetes能够智能的判断升级中断之前的状态，然后紧接着继续执行升级。当然，也可以进行退出，命令如下所示：
1 [root@k8s-master rc]# kubectl rolling-update myweb myweb2 --update-period=10s --rollback 2 Setting "myweb" replicas to 2 3 Continuing update with existing controller myweb. 4 Scaling up myweb from 2 to 2, scaling down myweb2 from 1 to 0 (keep 2 pods available, don‘t exceed 3 pods) 5 Scaling myweb2 down to 0 6 Update succeeded. Deleting myweb2 7 replicationcontroller "myweb" rolling updated to "myweb2" 8 [root@k8s-master rc]#