kind多节点集群

在 kind集群(Docker in Docker) 可以部署多个集群，不过默认部署的集群都是单个节点的，看上去和 minikube 并没有太大差别，只能验证应用功能，而不能实际演练Kubernetes集群的功能。

kind的真正功能在于部署多节点集群(multi-node clusters)，这种部署提供了完整的集群模拟。

kind官方网站提供了 kind-example-config ，可以通过以下命令创建一个定制的集群:

kind create cluster --config kind-example-config.yaml

多节点集群实践

• 配置3个管控节点，5个工作节点的集群配置文件如下：

kind-config.yaml

# a cluster with 3 control-plane nodes and 5 workers
kind: Cluster
apiVersion: kind.x-k8s.io/v1alpha4
nodes:
- role: control-plane
- role: control-plane
- role: control-plane
- role: worker
- role: worker
- role: worker
- role: worker
- role: worker

• 执行创建集群，集群命名为 dev

  kind create cluster --name dev --config kind-config.yaml
  

• 完成以后检查集群:

  kubectl --context kind-dev get nodes
  

可以看到创建的集群节点如下:

NAME                 STATUS   ROLES    AGE     VERSION
dev-control-plane    Ready    master   2m15s   v1.18.2
dev-control-plane2   Ready    master   99s     v1.18.2
dev-control-plane3   Ready    master   49s     v1.18.2
dev-worker           Ready    <none>   32s     v1.18.2
dev-worker2          Ready    <none>   32s     v1.18.2
dev-worker3          Ready    <none>   32s     v1.18.2
dev-worker4          Ready    <none>   29s     v1.18.2
dev-worker5          Ready    <none>   32s     v1.18.2

通过 docker ps 命令可以看到，kind还多启动了一个容器来运行haproxy，以提供外部访问这个集群的apiserver的负载均衡分发，例如以下案例中端口 33401 就是访问该集群的apiserver端口(见第一行):

#docker ps
CONTAINER ID        IMAGE                          COMMAND                  CREATED             STATUS              PORTS                       NAMES
48309bdd8439        kindest/haproxy:2.1.1-alpine   "/docker-entrypoint.…"   42 hours ago        Up 42 hours         127.0.0.1:33401->6443/tcp   dev-external-load-balancer
b0f8267aa09a        kindest/node:v1.18.2           "/usr/local/bin/entr…"   42 hours ago        Up 42 hours                                     dev-worker2
524c7b50790b        kindest/node:v1.18.2           "/usr/local/bin/entr…"   42 hours ago        Up 42 hours                                     dev-worker
9790bbe67900        kindest/node:v1.18.2           "/usr/local/bin/entr…"   42 hours ago        Up 42 hours         127.0.0.1:21003->6443/tcp   dev-control-plane
21c86c9750dd        kindest/node:v1.18.2           "/usr/local/bin/entr…"   42 hours ago        Up 42 hours         127.0.0.1:32351->6443/tcp   dev-control-plane3
8abd58281f54        kindest/node:v1.18.2           "/usr/local/bin/entr…"   42 hours ago        Up 42 hours                                     dev-worker5
634df18e345e        kindest/node:v1.18.2           "/usr/local/bin/entr…"   42 hours ago        Up 42 hours                                     dev-worker4
06d46f9ff05e        kindest/node:v1.18.2           "/usr/local/bin/entr…"   42 hours ago        Up 42 hours         127.0.0.1:14861->6443/tcp   dev-control-plane2
a8f88cb59b1b        kindest/node:v1.18.2           "/usr/local/bin/entr…"   42 hours ago        Up 42 hours                                     dev-worker3

所以，在运行 kind 的服务器上执行:

kubectl cluster-info

可以看到如下输出:

Kubernetes master is running at https://127.0.0.1:33401
KubeDNS is running at https://127.0.0.1:33401/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy

To further debug and diagnose cluster problems, use 'kubectl cluster-info dump'.