Kubernetes(k8s)的RC(Replication Controller)副本控制器
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。
[ ~]# kubectl get nods the server doesn‘t have a resource type "nods" [ ~]# kubectl get node NAME STATUS AGE k8s-master Ready 6d k8s-node2 Ready 6d k8s-node3 Ready 6d [ ~]# kubectl get nodes NAME STATUS AGE k8s-master Ready 6d k8s-node2 Ready 6d k8s-node3 Ready 6d [ ~]# clear [ ~]# kubectl get componentstatus NAME STATUS MESSAGE ERROR scheduler Healthy ok controller-manager Healthy ok etcd-0 Healthy {"health":"true"} [ ~]# clear [ ~]# cd k8s/ [master k8s]# ls pod [master k8s]# mkdir rc [ k8s]# cd rc/ [master rc]# ls [master rc]# vim nginx_rc_yaml [ rc]# kubectl create -f nginx_rc_yaml replicationcontroller "myweb" created [ rc]# kubectl get rc NAME DESIRED CURRENT READY AGE myweb 2 2 1 13s [ rc]#
创建nginx_rc_yaml配置文件,配置内容,如下所示。
# 声明api的版本。 apiVersion: v1 # kind代表资源的类型,资源是ReplicationController。 kind: ReplicationController # 资源叫什么名字,是在其属性metadata里面的。 metadata: # 第一个属性name的值是myweb,即ReplicationController的名字就叫做myweb。 name: myweb # spec是详细,详细里面定义了一个容器。 spec: # 声明副本数量是2,代表了RC会启动两个相同的Pod。 replicas: 2 # 选择器。 selector: app: myweb # 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。 # 没有指定名称,RC会随机生成一个名称。 template: # 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。 metadata: # 给Pod贴上了一个标签,标签是app: web,标签是有一定的作用的。 labels: app: myweb # spec是详细,详细里面定义了一个容器。 spec: # 定义一个容器,可以声明多个容器的。 containers: # 容器的名称叫做myweb - name: myweb # 使用了什么镜像,可以使用官方公有的,也可以使用私有的。 image: 192.168.110.133:5000/nginx:1.13 # ports定义容器的端口。 ports: # 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。 - 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。
[ rc]# kubectl get rc NAME DESIRED CURRENT READY AGE myweb 2 2 1 7m [ rc]# kubectl get pods NAME READY STATUS RESTARTS AGE myweb-0hqc5 0/1 ImagePullBackOff 0 8m myweb-2np4k 1/1 Running 0 8m nginx 1/1 Running 1 3d test1 0/1 ImagePullBackOff 0 2d test2 2/2 Running 1 2d test4 1/1 Running 0 2d [ 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。
[ ~]# kubectl get rc NAME DESIRED CURRENT READY AGE myweb 2 2 1 17m [ ~]# kubectl get pod -o wide NAME READY STATUS RESTARTS AGE IP NODE myweb-8cp7h 0/1 ImagePullBackOff 0 5m 172.16.85.3 k8s-master myweb-qcgjl 1/1 Running 1 14m 172.16.5.2 k8s-node2 nginx 1/1 Running 2 3d 172.16.38.3 k8s-node3 test1 0/1 ImagePullBackOff 0 2d 172.16.85.2 k8s-master test2 2/2 Running 3 2d 172.16.38.2 k8s-node3 test4 1/1 Running 1 2d 172.16.5.3 k8s-node2 [ ~]# kubectl delete rc myweb replicationcontroller "myweb" deleted [ ~]# kubectl get rc No resources found. [ ~]#
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。
[ ~]# docker images REPOSITORY TAG IMAGE ID CREATED SIZE docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB 192.168.110.133:5000/nginx 1.13 ae513a47849c 2 years ago 109 MB docker.io/nginx 1.13 ae513a47849c 2 years ago 109 MB 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB docker.io/tianyebj/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB [ ~]#
而我的RC的yaml的配置文件,如下所示:
Version: v1 kind: ReplicationController metadata: name: myweb spec: replicas: 2 selector: app: myweb template: metadata: labels: app: myweb spec: containers: - name: myweb image: 192.168.110.133:5000/nginx:1.13 # imagePullPolicy: IfNotPresent ports: - 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。
[ ~]# docker images REPOSITORY TAG IMAGE ID CREATED SIZE docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB docker.io/registry latest 708bc6af7e5e 4 months ago 25.8 MB 192.168.110.133:5000/ngnix 1.13 ae513a47849c 2 years ago 109 MB registry.access.redhat.com/rhel7/pod-infrastructure latest 99965fb98423 2 years ago 209 MB 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB [ ~]#
报错信息,也贴一下吧,方便自己以后使用,如下所示:
[ ~]# kubectl describe pod myweb-qwgsf Name: myweb-qwgsf Namespace: default Node: k8s-master/192.168.110.133 Start Time: Thu, 11 Jun 2020 17:21:45 +0800 Labels: app=myweb Status: Pending IP: 172.16.85.2 Controllers: ReplicationController/myweb Containers: myweb: Container ID: Image: 192.168.110.133:5000/nginx:1.13 Image ID: Port: 80/TCP State: Waiting Reason: ErrImagePull Ready: False Restart Count: 0 Volume Mounts: <none> Environment Variables: <none> Conditions: Type Status Initialized True Ready False PodScheduled True No volumes. QoS Class: BestEffort Tolerations: <none> Events: FirstSeen LastSeen Count From SubObjectPath Type Reason Message --------- -------- ----- ---- ------------- -------- ------ ------- 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. 12m 12m 1 {default-scheduler } Normal Scheduled Successfully assigned myweb-qwgsf to k8s-master 12m 6m 6 {kubelet k8s-master} spec.containers{myweb} Normal Pulling pulling image "192.168.110.133:5000/nginx:1.13" 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 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" 12m 4m 31 {kubelet k8s-master} spec.containers{myweb} Normal BackOff Back-off pulling image "192.168.110.133:5000/nginx:1.13" 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\"" 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. 31s 20s 2 {kubelet k8s-master} spec.containers{myweb} Normal Pulling pulling image "192.168.110.133:5000/nginx:1.13" 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 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" 30s 8s 2 {kubelet k8s-master} spec.containers{myweb} Normal BackOff Back-off pulling image "192.168.110.133:5000/nginx:1.13" 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。
[ ~]# docker images REPOSITORY TAG IMAGE ID CREATED SIZE docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB docker.io/registry latest 708bc6af7e5e 4 months ago 25.8 MB 192.168.110.133:5000/ngnix 1.13 ae513a47849c 2 years ago 109 MB registry.access.redhat.com/rhel7/pod-infrastructure latest 99965fb98423 2 years ago 209 MB 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB [ ~]# docker ps CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES [ ~]# docker ps -a CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES a27987d97039 registry "/entrypoint.sh /e..." 5 days ago Exited (2) 3 days ago registry ee95778bd5d9 busybox "sh" 6 days ago Exited (127) 6 days ago friendly_payne 6d459781a3e5 busybox "sh" 6 days ago Exited (137) 5 days ago gracious_nightingale [ ~]# docker rmi -f ae513a47849c Untagged: 192.168.110.133:5000/ngnix:1.13 Untagged: 192.168.110.133:5000/:e4f0474a75c510f40b37b6b7dc2516241ffa8bde5a442bde3d372c9519c84d90 Deleted: sha256:ae513a47849c895a155ddfb868d6ba247f60240ec8495482eca74c4a2c13a881 Deleted: sha256:160a8bd939a9421818f499ba4fbfaca3dd5c86ad7a6b97b6889149fd39bd91dd Deleted: sha256:f246685cc80c2faa655ba1ec9f0a35d44e52b6f83863dc16f46c5bca149bfefc Deleted: sha256:d626a8ad97a1f9c1f2c4db3814751ada64f60aed927764a3f994fcd88363b659 [ ~]# docker images REPOSITORY TAG IMAGE ID CREATED SIZE docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB docker.io/registry latest 708bc6af7e5e 4 months ago 25.8 MB registry.access.redhat.com/rhel7/pod-infrastructure latest 99965fb98423 2 years ago 209 MB 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB [ ~]# docker images REPOSITORY TAG IMAGE ID CREATED SIZE docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB docker.io/registry latest 708bc6af7e5e 4 months ago 25.8 MB registry.access.redhat.com/rhel7/pod-infrastructure latest 99965fb98423 2 years ago 209 MB 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB [ ~]#
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。
[ ~]# free -h total used free shared buff/cache available Mem: 2.2G 880M 453M 12M 953M 1.1G Swap: 2.0G 0B 2.0G [ ~]# kubectl get node NAME STATUS AGE k8s-master Ready 6d k8s-node2 Ready 6d k8s-node3 Ready 6d [ ~]# kubectl get componentstatus NAME STATUS MESSAGE ERROR scheduler Healthy ok controller-manager Healthy ok etcd-0 Healthy {"health":"true"} [ ~]# kubectl get rc NAME DESIRED CURRENT READY AGE myweb 2 2 2 16m [ ~]# docker images REPOSITORY TAG IMAGE ID CREATED SIZE docker.io/nginx latest 2622e6cca7eb 41 hours ago 132 MB docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB docker.io/registry latest 708bc6af7e5e 4 months ago 25.8 MB 192.168.110.133:5000/nginx 1.13 ae513a47849c 2 years ago 109 MB docker.io/nginx 1.13 ae513a47849c 2 years ago 109 MB registry.access.redhat.com/rhel7/pod-infrastructure latest 99965fb98423 2 years ago 209 MB 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB [ ~]# kubectl get node -o wide NAME STATUS AGE EXTERNAL-IP k8s-master Ready 6d <none> k8s-node2 Ready 6d <none> k8s-node3 Ready 6d <none> [ ~]# kubectl get pod -o wide NAME READY STATUS RESTARTS AGE IP NODE myweb-2h8b1 1/1 Running 1 17m 172.16.85.2 k8s-master myweb-lfkmp 1/1 Running 1 17m 172.16.5.2 k8s-node2 test4 1/1 Running 1 13m 172.16.38.2 k8s-node3 [ ~]#
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。
[ rc]# kubectl describe pod myweb-2h8b1 Name: myweb-2h8b1 Namespace: default Node: k8s-master/192.168.110.133 Start Time: Thu, 11 Jun 2020 17:51:06 +0800 Labels: app=myweb Status: Running IP: 172.16.85.2 Controllers: ReplicationController/myweb Containers: myweb: Container ID: docker://27a9e6dfb65be540bb50c98d820a5b773c0ed01d09d2350baf6027cdf9e22257 Image: 192.168.110.133:5000/nginx:1.13 Image ID: docker-pullable://docker.io/:b1d09e9718890e6ebbbd2bc319ef1611559e30ce1b6f56b2e3b479d9da51dc35 Port: 80/TCP State: Running Started: Thu, 11 Jun 2020 18:06:34 +0800 Last State: Terminated Reason: Completed Exit Code: 0 Started: Thu, 11 Jun 2020 18:01:41 +0800 Finished: Thu, 11 Jun 2020 18:05:42 +0800 Ready: True Restart Count: 1 Volume Mounts: <none> Environment Variables: <none> Conditions: Type Status Initialized True Ready True PodScheduled True No volumes. QoS Class: BestEffort Tolerations: <none> No events. [ rc]#
通过查看RC的标签。
[ rc]# kubectl get rc -o wide NAME DESIRED CURRENT READY AGE CONTAINER(S) IMAGE(S) SELECTOR myweb 2 2 2 3h myweb 192.168.110.133:5000/nginx:1.13 app=myweb [ 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。
[ rc]# kubectl get all NAME DESIRED CURRENT READY AGE rc/myweb 2 2 2 3h NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE svc/kubernetes 10.254.0.1 <none> 443/TCP 6d NAME READY STATUS RESTARTS AGE po/myweb-2h8b1 1/1 Running 1 3h po/myweb-lfkmp 1/1 Running 1 3h po/test4 1/1 Running 1 3h [master rc]# kubectl edit pod test4 pod "test4" edited [ rc]# kubectl get all NAME DESIRED CURRENT READY AGE rc/myweb 2 2 2 3h NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE svc/kubernetes 10.254.0.1 <none> 443/TCP 6d NAME READY STATUS RESTARTS AGE po/myweb-lfkmp 1/1 Running 1 3h po/test4 1/1 Running 1 3h [ 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。
[master rc]# cp nginx_rc.yaml nginx_rc2.yaml [master rc]# ls nginx_rc2.yaml nginx_rc.yaml [master rc]# vim nginx_rc2.yaml [ 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 [ ~]# 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 [ ~]# 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,[ ~]# 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/
[ ~]# docker pull docker.io/nginx:1.15 Trying to pull repository docker.io/library/nginx ... sha256:23b4dcdf0d34d4a129755fc6f52e1c6e23bb34ea011b315d87e193033bcd1b68: Pulling from docker.io/library/nginx 743f2d6c1f65: Pull complete 6bfc4ec4420a: Pull complete 688a776db95f: Pull complete Digest: sha256:23b4dcdf0d34d4a129755fc6f52e1c6e23bb34ea011b315d87e193033bcd1b68 Status: Downloaded newer image for docker.io/nginx:1.15 [ ~]#
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。
[master rc]# docker images REPOSITORY TAG IMAGE ID CREATED SIZE docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB docker.io/registry latest 708bc6af7e5e 4 months ago 25.8 MB docker.io/nginx 1.15 53f3fd8007f7 13 months ago 109 MB 192.168.110.133:5000/nginx 1.13 ae513a47849c 2 years ago 109 MB docker.io/nginx 1.13 ae513a47849c 2 years ago 109 MB registry.access.redhat.com/rhel7/pod-infrastructure latest 99965fb98423 2 years ago 209 MB 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB [ rc]# docker tag docker.io/nginx:1.15 192.168.110.133:5000/nginx:1.15 [ rc]# docker push 192.168.110.133:5000/nginx:1.15 The push refers to a repository [192.168.110.133:5000/nginx] Put http://192.168.110.133:5000/v1/repositories/nginx/: dial tcp 192.168.110.133:5000: connect: connection refused [master rc]# docker images REPOSITORY TAG IMAGE ID CREATED SIZE docker.io/busybox latest 1c35c4412082 8 days ago 1.22 MB docker.io/registry latest 708bc6af7e5e 4 months ago 25.8 MB docker.io/nginx 1.15 53f3fd8007f7 13 months ago 109 MB 192.168.110.133:5000/nginx 1.15 53f3fd8007f7 13 months ago 109 MB docker.io/nginx 1.13 ae513a47849c 2 years ago 109 MB 192.168.110.133:5000/nginx 1.13 ae513a47849c 2 years ago 109 MB registry.access.redhat.com/rhel7/pod-infrastructure latest 99965fb98423 2 years ago 209 MB 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB [ 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:
[node3 docker]# docker images REPOSITORY TAG IMAGE ID CREATED SIZE docker.io/busybox latest 1c35c4412082 9 days ago 1.22 MB 192.168.110.133:5000/nginx 1.15 53f3fd8007f7 13 months ago 109 MB docker.io/nginx 1.15 53f3fd8007f7 13 months ago 109 MB 192.168.110.133:5000/nginx 1.13 ae513a47849c 2 years ago 109 MB docker.io/tianyebj/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB 192.168.110.133:5000/pod-infrastructure latest 34d3450d733b 3 years ago 205 MB [ 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。
[ ~]# docker version Client: Version: 1.13.1 API version: 1.26 Package version: docker-1.13.1-161.git64e9980.el7_8.x86_64 Go version: go1.10.3 Git commit: 64e9980/1.13.1 Built: Tue Apr 28 14:43:01 2020 OS/Arch: linux/amd64 Server: Version: 1.13.1 API version: 1.26 (minimum version 1.12) Package version: docker-1.13.1-161.git64e9980.el7_8.x86_64 Go version: go1.10.3 Git commit: 64e9980/1.13.1 Built: Tue Apr 28 14:43:01 2020 OS/Arch: linux/amd64 Experimental: false [ ~]#
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。
[ ~]# cat /etc/docker/daemon.json_bak { "insecure-registries":["192.168.110.133:5000"]} [ ~]# cd /etc/docker/ [node3 docker]# ls certs.d daemon.json_20200612 daemon.json_bak key.json seccomp.json [ 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。
[ docker]# cat /etc/sysconfig/docker # /etc/sysconfig/docker # Modify these options if you want to change the way the docker daemon runs # OPTIONS=‘--selinux-enabled --log-driver=journald --signature-verification=false‘ # 信任私有仓库,镜像加速 OPTIONS=‘--selinux-enabled --log-driver=journald --signature-verification=false --registry-mirror=https://registry.docker-cn.com --insecure-registry=192.168.110.133:5000‘ if [ -z "${DOCKER_CERT_PATH}" ]; then DOCKER_CERT_PATH=/etc/docker fi # Do not add registries in this file anymore. Use /etc/containers/registries.conf # instead. For more information reference the registries.conf(5) man page. # Location used for temporary files, such as those created by # docker load and build operations. Default is /var/lib/docker/tmp # Can be overriden by setting the following environment variable. # DOCKER_TMPDIR=/var/tmp # Controls the /etc/cron.daily/docker-logrotate cron job status. # To disable, uncomment the line below. # LOGROTATE=false # docker-latest daemon can be used by starting the docker-latest unitfile. # To use docker-latest client, uncomment below lines #DOCKERBINARY=/usr/bin/docker-latest #DOCKERDBINARY=/usr/bin/dockerd-latest #DOCKER_CONTAINERD_BINARY=/usr/bin/docker-containerd-latest #DOCKER_CONTAINERD_SHIM_BINARY=/usr/bin/docker-containerd-shim-latest [ 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:
[ rc]# kubectl rolling-update myweb -f nginx_rc2.yaml --update-period=10s Created myweb2 Scaling up myweb2 from 0 to 2, scaling down myweb from 2 to 0 (keep 2 pods available, don‘t exceed 3 pods) Scaling myweb2 up to 1 Scaling myweb down to 1 Scaling myweb2 up to 2 Scaling myweb down to 0 Update succeeded. Deleting myweb replicationcontroller "myweb" rolling updated to "myweb2" [ 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。
[ ~]# kubectl get pod -o wide NAME READY STATUS RESTARTS AGE IP NODE myweb2-f5400 1/1 Running 0 15s 172.16.38.3 k8s-node3 myweb2-mg9sk 1/1 Running 0 26s 172.16.85.2 k8s-master [ ~]#
升级之后还可以进行回滚,如下所示:
[ rc]# kubectl rolling-update myweb2 -f nginx_rc.yaml --update-period=10s Created myweb Scaling up myweb from 0 to 2, scaling down myweb2 from 2 to 0 (keep 2 pods available, don‘t exceed 3 pods) Scaling myweb up to 1 Scaling myweb2 down to 1 Scaling myweb up to 2 Scaling myweb2 down to 0 Update succeeded. Deleting myweb2 replicationcontroller "myweb2" rolling updated to "myweb" [ 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:
[ rc]# kubectl rolling-update myweb myweb2 --update-period=10s --rollback Setting "myweb" replicas to 2 Continuing update with existing controller myweb. Scaling up myweb from 2 to 2, scaling down myweb2 from 1 to 0 (keep 2 pods available, don‘t exceed 3 pods) Scaling myweb2 down to 0 Update succeeded. Deleting myweb2 replicationcontroller "myweb" rolling updated to "myweb2" [ rc]#