第十四章 二进制部署k8s集群的平滑升级

1、软件包下载

去github上下载较新的Kubernetes软件包https://github.com/

2、升级说明

• 升级包括master节点升级和node节点的升级，本章升级至v1.15.12；
• Master节点的服务包括：apiserver、controller-manager、kube-scheduler；
• Node节点的服务包括：kubelet和kube-proxy；
• 由于apiserver被nginx代理，所以在升级的时候需要操作操作nginx注释升级节点，避免带来无法访问的情况；
• 我们的master节点和node都是在同一个集群服务器上，所以一起进行操作；

3、确定节点升级顺序

查看节点信息

[root@hdss7-21 ~]# kubectl get node
NAME                STATUS   ROLES    AGE   VERSION
hdss7-21.host.com   Ready    <none>   14d   v1.14.10
hdss7-22.host.com   Ready    <none>   14d   v1.14.10

查看pod分布状态，尽量选择较少pod的节点先进行迁移

[root@hdss7-21 ~]# kubectl get pod -o wide -n kube-system
NAME                                   READY   STATUS    RESTARTS   AGE   IP           NODE                NOMINATED NODE   READINESS GATES
coredns-64f49f5655-smzzz               1/1     Running   6          8d    172.7.21.4   hdss7-21.host.com   <none>           <none>
kubernetes-dashboard-99ff79fcd-khl8z   1/1     Running   2          4d    172.7.22.4   hdss7-22.host.com   <none>           <none>
traefik-ingress-2svq6                  1/1     Running   3          5d    172.7.21.5   hdss7-21.host.com   <none>           <none>
traefik-ingress-rcd28                  1/1     Running   3          5d    172.7.22.3   hdss7-22.host.com   <none>           <none>

由于分布差不多，我们选择先升级10.4.7.21服务器上的节点

4、修改代理nginx配置

在10.4.7.21和22上都操作，以21为例

注释apiserver升级节点的服务器

[root@hdss7-11 ~]# vim /etc/nginx/nginx.conf
    upstream kube-apiserver {
#        server 10.4.7.21:6443     max_fails=3 fail_timeout=30s;
        server 10.4.7.22:6443     max_fails=3 fail_timeout=30s;
    }
[root@hdss7-11 ~]# nginx -t
nginx: the configuration file /etc/nginx/nginx.conf syntax is ok
nginx: configuration file /etc/nginx/nginx.conf test is successful
[root@hdss7-11 ~]# nginx -s reload

5、删除第一个节点

将节点调成不可调度状态

[root@hdss7-21 ~]#  kubectl cordon hdss7-21.host.com
node/hdss7-21.host.com cordoned

当节点设置成不可调度状态之后，新启动的 pod 不会调度到此节点上，但是该节点上正在运行的 Pod 将不会被影响。

驱逐节点上的pod

[root@hdss7-21 ~]# kubectl drain hdss7-21.host.com --delete-local-data --ignore-daemonsets --force
node/hdss7-21.host.com already cordoned
WARNING: ignoring DaemonSet-managed Pods: default/nginx-ds-2rj9d, kube-system/traefik-ingress-2svq6
evicting pod "coredns-64f49f5655-smzzz"
evicting pod "nginx-dp-86678bb55c-tklvc"
pod/nginx-dp-86678bb55c-tklvc evicted
pod/coredns-64f49f5655-smzzz evicted
node/hdss7-21.host.com evicted

注释：

--delete-local-data 即使pod使用了emptyDir也删除 --ignore-daemonsets 忽略deamonset控制器的pod，如果不忽略，daemonset控制器控制的pod被删除后可能马上又在此节点上启动起来,会成为死循环； --force 不加force参数只会删除该NODE上由ReplicationController, ReplicaSet, DaemonSet,StatefulSet or Job创建的Pod，加了后还会删除'裸奔的pod'(没有绑定到任何replication controller)

再次查看pod分布

[root@hdss7-21 ~]# kubectl get pod -o wide -n kube-system
NAME                                   READY   STATUS    RESTARTS   AGE   IP           NODE                NOMINATED NODE   READINESS GATES
coredns-64f49f5655-n2pl7               1/1     Running   0          39s   172.7.22.7   hdss7-22.host.com   <none>           <none>
kubernetes-dashboard-99ff79fcd-khl8z   1/1     Running   2          4d    172.7.22.4   hdss7-22.host.com   <none>           <none>
traefik-ingress-2svq6                  1/1     Running   3          5d    172.7.21.5   hdss7-21.host.com   <none>           <none>
traefik-ingress-rcd28                  1/1     Running   3          5d    172.7.22.3   hdss7-22.host.com   <none>           <none>

除了daemonset的pod已被移动到10.4.7.22接节点上

测试重新启动的coredns是否生效

[root@hdss7-21 ~]# dig -t A nginx-dp.default.svc.cluster.local @192.168.0.2 +short
192.168.191.8

删除节点

[root@hdss7-21 ~]# kubectl delete node hdss7-21.host.com
node "hdss7-21.host.com" deleted

6、升级第一个节点

配置新版本

[root@hdss7-21 ~]# cd /opt/src/
[root@hdss7-21 src]# tar -zxvf kubernetes-server-linux-amd64-v1.15.12.tar.gz
[root@hdss7-21 src]# mv kubernetes /opt/kubernetes-v1.15.12
[root@hdss7-21 src]# cd /opt/kubernetes-v1.15.12/
[root@hdss7-21 kubernetes-v1.15.12]# rm -f kubernetes-src.tar.gz
[root@hdss7-21 kubernetes-v1.15.12]# cd server/bin/
[root@hdss7-21 bin]# rm -f *.tar *_tag
[root@hdss7-21 bin]# ll
总用量 677292
-rwxr-xr-x 1 root root  50581504 6月  16 21:10 apiextensions-apiserver
-rwxr-xr-x 1 root root  44638208 6月  16 21:10 kubeadm
-rwxr-xr-x 1 root root  48525312 6月  16 21:10 kube-aggregator
-rwxr-xr-x 1 root root 122097664 6月  16 21:10 kube-apiserver
-rwxr-xr-x 1 root root 116301824 6月  16 21:10 kube-controller-manager
-rwxr-xr-x 1 root root  46419968 6月  16 21:10 kubectl
-rwxr-xr-x 1 root root  54980712 6月  16 21:10 kubectl-convert
-rwxr-xr-x 1 root root 118151728 6月  16 21:10 kubelet
-rwxr-xr-x 1 root root  43139072 6月  16 21:10 kube-proxy
-rwxr-xr-x 1 root root  47112192 6月  16 21:10 kube-scheduler
-rwxr-xr-x 1 root root   1593344 6月  16 21:10 mounter
[root@hdss7-21 bin]# mkdir certs
[root@hdss7-21 bin]# mkdir /opt/kubernetes-v1.15.12/conf

拷贝证书

[root@hdss7-21 bin]# cp /opt/kubernetes/server/bin/certs/* certs/
[root@hdss7-21 bin]# ls certs/
apiserver-key.pem  ca-key.pem  client-key.pem  kubelet-key.pem  kube-proxy-client-key.pem
apiserver.pem      ca.pem      client.pem      kubelet.pem      kube-proxy-client.pem

拷贝服务启动脚本

[root@hdss7-21 bin]# cp /opt/kubernetes/server/bin/*.sh .
[root@hdss7-21 bin]# ls
apiextensions-apiserver  kube-apiserver                      kubectl             kube-proxy                 mounter
certs                    kube-apiserver-startup.sh           kubectl-convert     kube-proxy-startup.sh
kubeadm                  kube-controller-manager             kubelet             kube-scheduler
kube-aggregator          kube-controller-manager-startup.sh  kubelet-startup.sh  kube-scheduler-startup.sh

拷贝配置文件

[root@hdss7-21 bin]# cp /opt/kubernetes/conf/* /opt/kubernetes-v1.15.12/conf/
[root@hdss7-21 bin]# ls /opt/kubernetes-v1.15.12/conf/
audit.yaml  k8s-node.yaml  kubelet.kubeconfig  kube-proxy.kubeconfig  nginx-ds.yaml

重新创建软连接

[root@hdss7-21 bin]# cd /opt/
[root@hdss7-21 opt]# ll
总用量 24
drwx--x--x 4 root root 4096 7月   2 21:50 containerd
lrwxrwxrwx 1 root root   16 6月   8 20:36 etcd -> /opt/etcd-v3.3.1
drwxr-xr-x 5 etcd etcd 4096 7月   9 21:25 etcd-v3.3.1
lrwxrwxrwx 1 root root   20 7月  17 19:40 flannel -> /opt/flannel-v0.11.0
drwxr-xr-x 3 root root 4096 7月  28 20:05 flannel-v0.11.0
lrwxrwxrwx 1 root root   23 7月  10 20:17 kubernetes -> /opt/kubernetes-v1.14.10
drwxr-xr-x 5 root root 4096 7月  10 20:36 kubernetes-v1.14.10
drwxr-xr-x 6 root root 4096 7月  28 22:18 kubernetes-v1.15.12
drwxr-xr-x 2 root root 4096 7月  28 22:10 src
[root@hdss7-21 opt]# rm -rf kubernetes
[root@hdss7-21 opt]# ln -s /opt/kubernetes-v1.15.12 /opt/kubernetes
[root@hdss7-21 opt]# ll
总用量 24
drwx--x--x 4 root root 4096 7月   2 21:50 containerd
lrwxrwxrwx 1 root root   16 6月   8 20:36 etcd -> /opt/etcd-v3.3.1
drwxr-xr-x 5 etcd etcd 4096 7月   9 21:25 etcd-v3.3.1
lrwxrwxrwx 1 root root   20 7月  17 19:40 flannel -> /opt/flannel-v0.11.0
drwxr-xr-x 3 root root 4096 7月  28 20:05 flannel-v0.11.0
lrwxrwxrwx 1 root root   23 7月  28 22:20 kubernetes -> /opt/kubernetes-v1.15.12
drwxr-xr-x 5 root root 4096 7月  10 20:36 kubernetes-v1.14.10
drwxr-xr-x 6 root root 4096 7月  28 22:18 kubernetes-v1.15.12
drwxr-xr-x 2 root root 4096 7月  28 22:10 src

7、重启节点服务

[root@hdss7-21 opt]# supervisorctl status
etcd-server-7-21                 RUNNING   pid 6296, uptime 0:16:14
flanneld-7-21                    RUNNING   pid 7042, uptime 0:13:14
kube-apiserver-7-21              RUNNING   pid 7165, uptime 0:12:24
kube-controller-manager-7-21     RUNNING   pid 4675, uptime 0:19:03
kube-kubelet-7-21                RUNNING   pid 7184, uptime 0:12:16
kube-proxy-7-21                  RUNNING   pid 4678, uptime 0:19:03
kube-scheduler-7-21              RUNNING   pid 4673, uptime 0:19:03

重启node节点服务

[root@hdss7-21 opt]# supervisorctl restart kube-kubelet-7-21
[root@hdss7-21 opt]# supervisorctl restart kube-proxy-7-21

查看版本

[root@hdss7-21 opt]# kubectl get node
NAME                STATUS   ROLES    AGE     VERSION
hdss7-21.host.com   Ready    <none>   4d22h   v1.15.12
hdss7-22.host.com   Ready    <none>   19d     v1.14.10

重启master节点服务

[root@hdss7-21 opt]# supervisorctl restart kube-apiserver-7-21
[root@hdss7-21 opt]# supervisorctl restart kube-controller-manager-7-21
[root@hdss7-21 opt]# supervisorctl restart kube-scheduler-7-21

注意重启过程中可以查看日志，确保启动无问题。

8、修改代理nginx配置

修改11和12配置

[root@hdss7-11 ~]# vim /etc/nginx/nginx.conf
    upstream kube-apiserver {
        server 10.4.7.21:6443     max_fails=3 fail_timeout=30s;
#        server 10.4.7.22:6443     max_fails=3 fail_timeout=30s;
    }
[root@hdss7-11 ~]# nginx -t
nginx: the configuration file /etc/nginx/nginx.conf syntax is ok
nginx: configuration file /etc/nginx/nginx.conf test is successful
[root@hdss7-11 ~]# nginx -s reload

9、升级第二个节点

按照同样的方法对其余节点进行升级

在10.4.7.22上操作

最后检查服务启动状态和node节点版本

[root@hdss7-22 ~]# supervisorctl status
etcd-server-7-22                 RUNNING   pid 1235, uptime 1:10:58
flanneld-7-22                    RUNNING   pid 1203, uptime 1:10:59
kube-apiserver-7-22              RUNNING   pid 25776, uptime 0:01:39
kube-controller-manager-7-22     RUNNING   pid 26009, uptime 0:01:09
kube-kubelet-7-22                RUNNING   pid 23925, uptime 0:06:08
kube-proxy-7-22                  RUNNING   pid 24142, uptime 0:05:38
kube-scheduler-7-22              RUNNING   pid 26190, uptime 0:00:38
[root@hdss7-22 ~]# kubectl get node
NAME                STATUS   ROLES    AGE     VERSION
hdss7-21.host.com   Ready    <none>   4d22h   v1.15.12
hdss7-22.host.com   Ready    <none>   6m12s   v1.15.12

10、修改nginx代理

将配置文件改为原来状态

[root@hdss7-11 ~]# vim /etc/nginx/nginx.conf
    upstream kube-apiserver {
        server 10.4.7.21:6443     max_fails=3 fail_timeout=30s;
        server 10.4.7.22:6443     max_fails=3 fail_timeout=30s;
    }
[root@hdss7-11 ~]# nginx -t
nginx: the configuration file /etc/nginx/nginx.conf syntax is ok
nginx: configuration file /etc/nginx/nginx.conf test is successful
[root@hdss7-11 ~]# nginx -s reload

11、测试操作平台

12、重新分配pod节点

查看pod信息得知目前大部分的pod在10.4.7.21上，这不符合性能最优状态，我们可以在dashbard面板上删除一个pod，通过scheduler的计算，会在另外一个（22上）负载较少的node节点上重新启动这个pod

如下，删除coredns的pod

正在重启状态

启动后的状态