Tanzu Kubernetes Grid基本操作

查看上文中建立的Cluster

在上一文中建立了管理和负载的集群，可以通过tkg命令查看

查看管理集群

使用tkg get management-cluster 命令可以查看管理集群

[root@hop-172 ~]# tkg get management-cluster 
 MANAGEMENT-CLUSTER-NAME            CONTEXT-NAME                                                           STATUS  
 tkg-mgmt-vsphere-20210207152959 *  tkg-mgmt-vsphere-20210207152959-admin@tkg-mgmt-vsphere-20210207152959  Success 

如果有多个管理集群，这里也会一并列出。
在多个管理集群中切换，使用tkg set management-cluster命令。

查看Tanzu Kubernetes clusters负载集群

使用tkg get cluster 可以查看在default namespace负载集群的情况，不在default的cluster需要列出–namespace 选项。

 [root@hop-172 ~]# tkg get cluster
 NAME              NAMESPACE  STATUS   CONTROLPLANE  WORKERS  KUBERNETES        ROLES  
 contour-test      default    running  1/1           3/3      v1.19.3+vmware.1  <none> 
 my-cluster-dev-1  default    running  1/1           1/1      v1.19.3+vmware.1  <none> 

如上图，已经创建了两个负载集群contour-test和my-cluster-dev-1，并列出其参数。
States一栏表述了现在集群的状态，定义如下：

• creating: 正在创建控制平面
• createStalled: 创建控制平面的过程已停止
• deleting: 集群正在删除中
• failed: 创建控制平面失败
• running: 控制平面已完全初始化
• updating: 集群正在发布更新或正在扩展节点
• updateFailed: 集群更新过程失败
• updateStalled: 群集更新过程已停止
• No status: 集群的创建尚未开始

如果群集处于停止状态stalled state，检查与外部registry的网络连接，确保目标平台上有足够的资源来完成操作以及DHCP正确发布了IPv4地址。

使用–include-management-cluster参数同时查看管理和负载集群。

[root@hop-172 ~]# tkg get cluster --include-management-cluster
 NAME                             NAMESPACE   STATUS   CONTROLPLANE  WORKERS  KUBERNETES        ROLES      
 contour-test                     default     running  1/1           3/3      v1.19.3+vmware.1  <none>     
 my-cluster-dev-1                 default     running  1/1           1/1      v1.19.3+vmware.1  <none>     
 tkg-mgmt-vsphere-20210207152959  tkg-system  running  1/1           1/1      v1.19.3+vmware.1  management 

导出Tanzu Kubernetes Cluster 配置细节

使用tkg get cluster --output json/yaml导出

 [root@hop-172 ~]# tkg get cluster --output yaml >1.yaml
[root@hop-172 ~]# cat 1.yaml 
- name: contour-test
  namespace: default
  status: running
  plan: ""
  controlplane: 1/1
  workers: 3/3
  kubernetes: v1.19.3+vmware.1
  roles: []
- name: my-cluster-dev-1
  namespace: default
  status: running
  plan: ""
  controlplane: 1/1
  workers: 1/1
  kubernetes: v1.19.3+vmware.1
  roles: []

上面的例子将Cluster的配置导入了文件1.yaml中。

切换到负载集群

获取Tanzu Kubernetes Cluster Credentials

运行tkg get credentials可以将集群的凭据自动添加到kubeconfig文件中，在运行时指定集群的名称。

[root@hop-172 ~]# tkg get credentials contour-test
Credentials of workload cluster 'contour-test' have been saved 
You can now access the cluster by running 'kubectl config use-context contour-test-admin@contour-test'

接入到负载集群

根据提示，使用kubectl config use-context contour-test-admin@contour-test命令便可以切换到contour-test集群

[root@hop-172 ~]# kubectl config use-context contour-test-admin@contour-test
Switched to context "contour-test-admin@contour-test".
[root@hop-172 ~]# kubectl get node
NAME                                STATUS   ROLES    AGE     VERSION
contour-test-control-plane-trnnk    Ready    master   5m41s   v1.19.3+vmware.1
contour-test-md-0-96dcc5c79-ft78t   Ready    <none>   4m38s   v1.19.3+vmware.1
contour-test-md-0-96dcc5c79-h7gzg   Ready    <none>   4m37s   v1.19.3+vmware.1
contour-test-md-0-96dcc5c79-l8gj5   Ready    <none>   4m43s   v1.19.3+vmware.1

在负载查看系统容器

使用kubectl get pod -A命令查看系统运行生成的Pod

[root@hop-172 ~]# kubectl get pod -A
NAMESPACE     NAME                                                       READY   STATUS    RESTARTS   AGE
kube-system   antrea-agent-8b9bk                                         2/2     Running   0          7m
kube-system   antrea-agent-hmt7q                                         2/2     Running   0          6m2s
kube-system   antrea-agent-k56qs                                         2/2     Running   0          5m56s
kube-system   antrea-agent-sdmnn                                         2/2     Running   0          5m57s
kube-system   antrea-controller-5d594c5cc7-j8tc7                         1/1     Running   0          7m2s
kube-system   coredns-5d6f7c958-5xwjk                                    1/1     Running   0          6m57s
kube-system   coredns-5d6f7c958-ljrtc                                    1/1     Running   0          6m57s
kube-system   etcd-contour-test-control-plane-trnnk                      1/1     Running   0          6m47s
kube-system   kube-apiserver-contour-test-control-plane-trnnk            1/1     Running   0          6m47s
kube-system   kube-controller-manager-contour-test-control-plane-trnnk   1/1     Running   0          6m47s
kube-system   kube-proxy-45n6k                                           1/1     Running   0          6m2s
kube-system   kube-proxy-7gnrw                                           1/1     Running   0          5m56s
kube-system   kube-proxy-s5pgb                                           1/1     Running   0          5m57s
kube-system   kube-proxy-vwvfn                                           1/1     Running   0          6m58s
kube-system   kube-scheduler-contour-test-control-plane-trnnk            1/1     Running   0          6m46s
kube-system   kube-vip-contour-test-control-plane-trnnk                  1/1     Running   0          6m46s
kube-system   vsphere-cloud-controller-manager-zq2s4                     1/1     Running   0          6m59s
kube-system   vsphere-csi-controller-5b6f54ccc5-hc7j8                    5/5     Running   0          7m2s
kube-system   vsphere-csi-node-95x9x                                     3/3     Running   0          6m2s
kube-system   vsphere-csi-node-js9tr                                     3/3     Running   0          5m57s
kube-system   vsphere-csi-node-krntq                                     3/3     Running   0          5m56s
kube-system   vsphere-csi-node-vrgwg                                     3/3     Running   0          7m

Antrea, CNI container networking interface
coredns, 内部 DNS
etcd, 存储 key-value 值
kube-apiserver, Kubernetes API 服务器
kube-proxy, Kubernetes network 代理
kube-scheduler, 调度及可用性
vsphere-cloud-controller-manager, Kubernetes cloud provider for vSphere
kube-vip, 为Cluster API server提供负载均衡服务
vsphere-csi-controller and vsphere-csi-node, CNI-container storage interface for vSphere

第三方接入系统

将凭据保存在单独的kubeconfig文件中（例如，将其分发给开发人员），需要指定–export-file选项。
如：tkg get credentials my-cluster --export-file my-cluster-credentials
在第三方设备使用的时候，只需要把文件my-cluster-credentials和环境变量KUBECONFIG匹配就可以
export KUBECONFIG=my-cluster-credentials

也可以将~/.kube/config这个文件拷贝到第三方设备到 ~/.kube/config，然后可以切换到所有context，没有.kube目录就建一个（/usr/local/bin下以及有kubectl和tkg，方法见Step by step：安装 Tanzu Kubernetes Grid）

[root@localhost ~]# cd .kube
-bash: cd: .kube: No such file or directory
[root@localhost ~]# mkdir .kube
[root@localhost ~]# cp config.bak  ./.kube/config
[root@localhost ~]# kubectl config get-contexts
CURRENT   NAME                                                                    CLUSTER                           AUTHINFO                                NAMESPACE
          contour-test-admin@contour-test                                         contour-test                      contour-test-admin                      
          my-cluster-dev-1-admin@my-cluster-dev-1                                 my-cluster-dev-1                  my-cluster-dev-1-admin                  
          tkg-mgmt-vsphere-20210207152959-admin@tkg-mgmt-vsphere-20210207152959   tkg-mgmt-vsphere-20210207152959   tkg-mgmt-vsphere-20210207152959-admin   
*         tkg-wld-admin@tkg-wld                                                   tkg-wld                           tkg-wld-admin                           
[root@localhost ~]# kubectl config use-context tkg-mgmt-vsphere-20210207152959-admin@tkg-mgmt-vsphere-20210207152959
Switched to context "tkg-mgmt-vsphere-20210207152959-admin@tkg-mgmt-vsphere-20210207152959".
[root@localhost ~]# kubectl get node
NAME                                                    STATUS   ROLES    AGE   VERSION
tkg-mgmt-vsphere-20210207152959-control-plane-2xdwx     Ready    master   10d   v1.19.3+vmware.1
tkg-mgmt-vsphere-20210207152959-md-0-78f475f9c8-6qmpl   Ready    <none>   10d   v1.19.3+vmware.1
[root@localhost ~]# kubectl config use-context tkg-wld-admin@tkg-wld                                                
Switched to context "tkg-wld-admin@tkg-wld".
[root@localhost ~]# kubectl get node 
NAME                            STATUS   ROLES    AGE   VERSION
tkg-wld-control-plane-vv7m9     Ready    master   13h   v1.19.3+vmware.1
tkg-wld-md-0-7cc4fb8888-vnzbd   Ready    <none>   13h   v1.19.3+vmware.1

以上