使用Kubeadm搭建Kubernetes(1.12.2)集群

Kubeadm是Kubernetes官方提供的用於快速安裝Kubernetes集群的工具，伴隨Kubernetes每個版本的發布都會同步更新，在2018年將進入GA狀態，說明離生產環境中使用的距離越來越近了。

使用Kubeadm搭建Kubernetes集群本來是件很簡單的事，但由於眾所周知的原因，在中國大陸是無法訪問 k8s.gcr.io的。這就使我們無法按照官方的教程來創建集群。而國內的教程參差不齊，大多也無法運行成功，我也是踩了很多坑，才部署成功，故在此分享出來。

准備

• 多台Ubuntu 16.04+、CentOS 7或HypriotOSv1.0.1+ 系統。

• 每台機器最少2GB內存，2CPUs。

• 集群中所有機器之間網絡連接正常。

• 打開相應的端口，詳見：Check required ports。

• 關閉防火牆和selinux。

  # 關閉防火牆
  systemctl stop firewalld
  systemctl disable firewalld
  
  # 禁用SELINUX
  setenforce 0
  
  vim /etc/selinux/config
  SELINUX=disabled
  

• 關閉系統的Swap，Kubernetes 1.8開始要求必須禁用Swap，如果不關閉，默認配置下kubelet將無法啟動。

  # 關閉系統的Swap方法如下:
  # 編輯`/etc/fstab`文件，注釋掉引用`swap`的行，保存並重啟后輸入:
  sudo swapoff -a
  

• 驗證Mac地址和product_uuid是否唯一。

  Kubernetes要求集群中所有機器具有不同的Mac地址、產品uuid、Hostname。可以使用如下命令查看：

  # UUID
  cat /sys/class/dmi/id/product_uuid
  
  # Mac地址
  ip link
  
  # Hostname
  cat /etc/hostname
  

在本示例中使用2台Ubuntu 18.04主機：

cat /etc/hosts

192.168.0.8 ubuntu1
192.168.0.7 ubuntu2

安裝Docker

Kubernetes從1.6開始使用CRI(Container Runtime Interface)容器運行時接口。默認的容器運行時仍然是Docker，是使用kubelet中內置dockershim CRI來實現的。

Docker的安裝可以參考之前的博客：Docker初體驗。

需要注意的是，Kubernetes 1.12已經針對Docker的1.11.1, 1.12.1, 1.13.1, 17.03, 17.06, 17.09, 18.06等版本做了驗證，最低支持的Docker版本是1.11.1，最高支持是18.06，而Docker最新版本已經是18.09了，故我們安裝時需要指定版本為18.06.1-ce：

sudo apt install docker-ce=18.06.1~ce~3-0~ubuntu

安裝kubeadm, kubelet 和 kubectl

部署之前，我們需要安裝一下三個包：

• kubeadm: 引導啟動k8s集群的命令行工具。

• kubelet: 在群集中所有節點上運行的核心組件, 用來執行如啟動pods和containers等操作。

• kubectl: 操作集群的命令行工具。

首先添加apt-key：

sudo apt update && sudo apt install -y apt-transport-https curl
curl -s https://mirrors.aliyun.com/kubernetes/apt/doc/apt-key.gpg | sudo apt-key add -

添加kubernetes源：

sudo vim /etc/apt/sources.list.d/kubernetes.list

deb https://mirrors.aliyun.com/kubernetes/apt/ kubernetes-xenial main

安裝：

sudo apt update
sudo apt install -y kubelet kubeadm kubectl
sudo apt-mark hold kubelet kubeadm kubectl

使用kubeadm創建一個單Master集群

初始化Master節點

K8s的控制面板組件運行在Master節點上，包括etcd和API server（Kubectl便是通過API server與k8s通信）。

在執行初始化之前，我們還有一下3點需要注意：

1. 選擇一個網絡插件，並檢查它是否需要在初始化Master時指定一些參數，比如我們可能需要根據選擇的插件來設置--pod-network-cidr參數。參考：Installing a pod network add-on。

2. kubeadm使用eth0的默認網絡接口（通常是內網IP）做為Master節點的advertise address，如果我們想使用不同的網絡接口，可以使用--apiserver-advertise-address=<ip-address>參數來設置。如果適應IPv6，則必須使用IPv6d的地址，如：--apiserver-advertise-address=fd00::101。

3. 由於國內的網絡問題，建議使用kubeadm config images pull來預先拉取初始化需要用到的鏡像，並檢查是否能連接到gcr.io的registries。

很明顯，在國內並不能訪問gcr.io，在上篇文章使用kubeadm搭建Kubernetes(1.10.2)集群（國內環境）中使用了打tag的方式，而這次，我們通過修改配置文件來拉實現。

在kubeadm v1.11+版本中，增加了一個kubeadm config print-default命令，可以讓我們方便的將kubeadm的默認配置打印到文件中：

kubeadm config print-default > kubeadm.conf 

然后我們修改kubeadm.conf中的鏡像倉儲地址：

sed -i "s/imageRepository: .*/imageRepository: registry.aliyuncs.com\/google_containers/g" kubeadm.conf

指定我們要的版本號，避免初始化時從https://dl.k8s.io/release/stable-1.12.txt讀取，可使用如下命令來設置：

sed -i "s/kubernetesVersion: .*/kubernetesVersion: v1.12.2/g" kubeadm.conf

現在我們可以使用--config參數指定kubeadm.conf文件來運行kubeadm的images pull的命令：

kubeadm config images pull --config kubeadm.conf

W1103 06:10:18.782958   23149 common.go:105] WARNING: Detected resource kinds that may not apply: [InitConfiguration MasterConfiguration JoinConfiguration NodeConfiguration]
[config] WARNING: Ignored YAML document with GroupVersionKind kubeadm.k8s.io/v1alpha3, Kind=JoinConfiguration
[config/images] Pulled registry.aliyuncs.com/google_containers/kube-apiserver:v1.12.2
[config/images] Pulled registry.aliyuncs.com/google_containers/kube-controller-manager:v1.12.2
[config/images] Pulled registry.aliyuncs.com/google_containers/kube-scheduler:v1.12.2
[config/images] Pulled registry.aliyuncs.com/google_containers/kube-proxy:v1.12.2
[config/images] Pulled registry.aliyuncs.com/google_containers/pause:3.1
[config/images] Pulled registry.aliyuncs.com/google_containers/etcd:3.2.24
[config/images] Pulled registry.aliyuncs.com/google_containers/coredns:1.2.2

可以看到，已經成功拉取了需要的鏡像。

但是，此處還有一個坑，基礎鏡像pause的拉取地址需要單獨設置，否則還是會從k8s.gcr.io來拉取，導致init的時候卡住，並最終失敗：

[init] this might take a minute or longer if the control plane images have to be pulled

Unfortunately, an error has occurred:
timed out waiting for the condition

This error is likely caused by:
- The kubelet is not running
- The kubelet is unhealthy due to a misconfiguration of the node in some way (required cgroups disabled)

If you are on a systemd-powered system, you can try to troubleshoot the error with the following commands:
- 'systemctl status kubelet'
- 'journalctl -xeu kubelet'

解決辦法有2種：

最簡單就是打一個k8s.gcr.io/pause:3.1的Tag:

docker tag registry.aliyuncs.com/google_containers/pause:3.1 k8s.gcr.io/pause:3.1

其次可以通過修改kubeadm.conf中的InitConfiguration的nodeRegistration:kubeletExtraArgs:pod-infra-container-image參數來設置基礎鏡像，大約在14行，修改后如下：

kind: InitConfiguration
nodeRegistration:
  kubeletExtraArgs:
    pod-infra-container-image: registry.aliyuncs.com/google_containers/pause:3.1

通常，我們在執行init命令時，可能還需要指定advertiseAddress、--pod-network-cidr等參數，但是由於我們這里使用kubeadm.conf配置文件來初始化，就不能在命令行中指定其他參數了，因此需要我們在kubeadm.conf來設置。

如下，我們修改kubeadm.conf中與--apiserver-advertise-address參數對應的advertiseAddress參數，我的虛擬機IP是:192.168.0.8，大家根據自己的實際情況來設置：

sed -i "s/advertiseAddress: .*/advertiseAddress: 192.168.0.8/g" kubeadm.conf

在本示例中，我使用的是Canal網絡插件，因此需要將--pod-network-cid設置為10.244.0.0/16，修改如下：

sed -i "s/podSubnet: .*/podSubnet: \"10.244.0.0\/16\"/g" kubeadm.conf

現在可以執行初始化命令了：

sudo kubeadm init --config kubeadm.conf

輸出如下：

W1109 17:01:47.071494   42929 common.go:105] WARNING: Detected resource kinds that may not apply: [InitConfiguration MasterConfiguration JoinConfiguration NodeConfiguration]
[config] WARNING: Ignored YAML document with GroupVersionKind kubeadm.k8s.io/v1alpha3, Kind=JoinConfiguration
[init] using Kubernetes version: v1.12.2
[preflight] running pre-flight checks
[preflight/images] Pulling images required for setting up a Kubernetes cluster
[preflight/images] This might take a minute or two, depending on the speed of your internet connection
[preflight/images] You can also perform this action in beforehand using 'kubeadm config images pull'
[kubelet] Writing kubelet environment file with flags to file "/var/lib/kubelet/kubeadm-flags.env"
[kubelet] Writing kubelet configuration to file "/var/lib/kubelet/config.yaml"
[preflight] Activating the kubelet service
[certificates] Generated ca certificate and key.
[certificates] Generated apiserver certificate and key.
[certificates] apiserver serving cert is signed for DNS names [ubuntu1 kubernetes kubernetes.default kubernetes.default.svc kubernetes.default.svc.cluster.local] and IPs [10.96.0.1 192.168.0.8]
[certificates] Generated apiserver-kubelet-client certificate and key.
[certificates] Generated front-proxy-ca certificate and key.
[certificates] Generated front-proxy-client certificate and key.
[certificates] Generated etcd/ca certificate and key.
[certificates] Generated etcd/server certificate and key.
[certificates] etcd/server serving cert is signed for DNS names [ubuntu1 localhost] and IPs [127.0.0.1 ::1]
[certificates] Generated apiserver-etcd-client certificate and key.
[certificates] Generated etcd/peer certificate and key.
[certificates] etcd/peer serving cert is signed for DNS names [ubuntu1 localhost] and IPs [192.168.0.8 127.0.0.1 ::1]
[certificates] Generated etcd/healthcheck-client certificate and key.
[certificates] valid certificates and keys now exist in "/etc/kubernetes/pki"
[certificates] Generated sa key and public key.
[kubeconfig] Wrote KubeConfig file to disk: "/etc/kubernetes/admin.conf"
[kubeconfig] Wrote KubeConfig file to disk: "/etc/kubernetes/kubelet.conf"
[kubeconfig] Wrote KubeConfig file to disk: "/etc/kubernetes/controller-manager.conf"
[kubeconfig] Wrote KubeConfig file to disk: "/etc/kubernetes/scheduler.conf"
[controlplane] wrote Static Pod manifest for component kube-apiserver to "/etc/kubernetes/manifests/kube-apiserver.yaml"
[controlplane] wrote Static Pod manifest for component kube-controller-manager to "/etc/kubernetes/manifests/kube-controller-manager.yaml"
[controlplane] wrote Static Pod manifest for component kube-scheduler to "/etc/kubernetes/manifests/kube-scheduler.yaml"
[etcd] Wrote Static Pod manifest for a local etcd instance to "/etc/kubernetes/manifests/etcd.yaml"
[init] waiting for the kubelet to boot up the control plane as Static Pods from directory "/etc/kubernetes/manifests" 
[init] this might take a minute or longer if the control plane images have to be pulled
[apiclient] All control plane components are healthy after 57.002438 seconds
[uploadconfig] storing the configuration used in ConfigMap "kubeadm-config" in the "kube-system" Namespace
[kubelet] Creating a ConfigMap "kubelet-config-1.12" in namespace kube-system with the configuration for the kubelets in the cluster
[markmaster] Marking the node ubuntu1 as master by adding the label "node-role.kubernetes.io/master=''"
[markmaster] Marking the node ubuntu1 as master by adding the taints [node-role.kubernetes.io/master:NoSchedule]
[patchnode] Uploading the CRI Socket information "/var/run/dockershim.sock" to the Node API object "ubuntu1" as an annotation
[bootstraptoken] using token: abcdef.0123456789abcdef
[bootstraptoken] configured RBAC rules to allow Node Bootstrap tokens to post CSRs in order for nodes to get long term certificate credentials
[bootstraptoken] configured RBAC rules to allow the csrapprover controller automatically approve CSRs from a Node Bootstrap Token
[bootstraptoken] configured RBAC rules to allow certificate rotation for all node client certificates in the cluster
[bootstraptoken] creating the "cluster-info" ConfigMap in the "kube-public" namespace
[addons] Applied essential addon: CoreDNS
[addons] Applied essential addon: kube-proxy

Your Kubernetes master has initialized successfully!

To start using your cluster, you need to run the following as a regular user:

  mkdir -p $HOME/.kube
  sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
  sudo chown $(id -u):$(id -g) $HOME/.kube/config

You should now deploy a pod network to the cluster.
Run "kubectl apply -f [podnetwork].yaml" with one of the options listed at:
  https://kubernetes.io/docs/concepts/cluster-administration/addons/

You can now join any number of machines by running the following on each node
as root:

  kubeadm join 192.168.0.8:6443 --token abcdef.0123456789abcdef --discovery-token-ca-cert-hash sha256:67ea537411822fe684d1ddb984802da62a4f22aa1c32fefe7c3404bb8f3f52e0

如果我們想使用非root用戶操作kubectl，可以使用以下命令，這也是kubeadm init輸出的一部分：

mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config

安裝網絡插件

為了讓Pods間可以相互通信，我們必須安裝一個網絡插件，並且必須在部署任何應用之前安裝，CoreDNS也是在網絡插件安裝之后才會啟動的。

網絡的插件完整列表，請參考 Networking and Network Policy。

在安裝之前，我們先查看一下當前Pods的狀態：

kubectl get pods --all-namespaces

# 輸出
NAMESPACE     NAME                              READY   STATUS    RESTARTS   AGE
kube-system   coredns-5c545769d8-j9vzw          0/1     Pending   0          110s
kube-system   coredns-5c545769d8-wqrlm          0/1     Pending   0          111s
kube-system   etcd-ubuntu1                      1/1     Running   0          75s
kube-system   kube-apiserver-ubuntu1            1/1     Running   0          87s
kube-system   kube-controller-manager-ubuntu1   1/1     Running   0          96s
kube-system   kube-proxy-snhqr                  1/1     Running   0          111s
kube-system   kube-scheduler-ubuntu1            1/1     Running   0          98s

如上，可以看到CoreDND的狀態是Pending，就是因為我們還沒有安裝網絡插件。

我是比較推薦的是Calico網絡插件，但是由於我的虛擬機網段是192.168.0.x，無法使用Calico網絡，所以使用了Canal網絡插件，它是Calico和Flannel的結合體，在上面kubeadm init的時候我們已經指定了--pod-network-cidr=10.244.0.0/16，這是Canal插件所要求的。

可使用如下命令命令來安裝Canal插件：

# 源地址：https://docs.projectcalico.org/v3.3/getting-started/kubernetes/installation/hosted/canal/rbac.yaml
kubectl apply -f http://mirror.faasx.com/k8s/canal/v3.3/rbac.yaml

# 源地址：https://docs.projectcalico.org/v3.3/getting-started/kubernetes/installation/hosted/canal/canal.yaml
# 只是將quay.io修改成了國內鏡像
kubectl apply -f http://mirror.faasx.com/k8s/canal/v3.3/canal.yaml

關於更多Canal的信息，可以查看Installing Calico for policy and flannel for networking。

稍等片刻，再使用kubectl get pods --all-namespaces命令來查看網絡插件的安裝情況：

NAMESPACE     NAME                              READY   STATUS    RESTARTS   AGE
kube-system   canal-frf6b                       3/3     Running   3          25m
kube-system   coredns-5c545769d8-j9vzw          1/1     Running   2          9h
kube-system   coredns-5c545769d8-wqrlm          1/1     Running   2          9h
kube-system   etcd-ubuntu1                      1/1     Running   1          9h
kube-system   kube-apiserver-ubuntu1            1/1     Running   1          9h
kube-system   kube-controller-manager-ubuntu1   1/1     Running   1          9h
kube-system   kube-proxy-snhqr                  1/1     Running   1          9h
kube-system   kube-scheduler-ubuntu1            1/1     Running   1          9h

如上，STATUS全部變為了Running，表示安裝成功，接下來就可以加入其他節點以及部署應用了。

Master隔離

默認情況下，由於安全原因，集群並不會將pods部署在Master節點上。但是在開發環境下，我們可能就只有一個Master節點，這時可以使用下面的命令來解除這個限制：

kubectl taint nodes --all node-role.kubernetes.io/master-

## 輸出
node/ubuntu1 untainted

加入工作節點

要為群集添加工作節點，需要為每台計算機執行以下操作：

• SSH到機器
• 成為root用戶，(如: sudo su -)
• 運行上面的kubeadm init命令輸出的：kubeadm join --token <token> <master-ip>:<master-port> --discovery-token-ca-cert-hash sha256:<hash>

如果我們忘記了Master節點的加入token，可以使用如下命令來查看：

kubeadm token list

# 輸出：
# TOKEN                     TTL       EXPIRES                USAGES                   DESCRIPTION   EXTRA GROUPS
# abcdef.0123456789abcdef   22h       2018-11-10T14:24:51Z   authentication,signing   <none>        system:bootstrappers:kubeadm:default-node-token

默認情況下，token的有效期是24小時，如果我們的token已經過期的話，可以使用以下命令重新生成：

kubeadm token create

# 輸出：
# 9w6mbu.3k2z7pprl3eaozk9

如果我們也沒有--discovery-token-ca-cert-hash的值，可以使用以下命令生成：

openssl x509 -pubkey -in /etc/kubernetes/pki/ca.crt | openssl rsa -pubin -outform der 2>/dev/null | openssl dgst -sha256 -hex | sed 's/^.* //'

# 輸出：
# 9fcb02a0f4ab216866f87986106437b7305474850f0de81b9ac9c36a468f7c67

現在，我們登錄到工作節點服務器，准備加入到集群。

但是還有最重要的一點就是，基礎鏡像pause需要單獨設置，否則還是會從k8s.gcr.io來拉取，我們可以使用類似Init時修改配置文件的方式來實現，不過，由於就這一個鏡像拉取有問題，我們可以簡單的打個tag：

docker pull registry.aliyuncs.com/google_containers/pause:3.1
docker tag registry.aliyuncs.com/google_containers/pause:3.1 k8s.gcr.io/pause:3.1

然后運行如下命令加入集群：

sudo kubeadm join 192.168.0.8:6443 --token abcdef.0123456789abcdef --discovery-token-ca-cert-hash sha256:67ea537411822fe684d1ddb984802da62a4f22aa1c32fefe7c3404bb8f3f52e0

輸出如下：

[preflight] running pre-flight checks
	[WARNING RequiredIPVSKernelModulesAvailable]: the IPVS proxier will not be used, because the following required kernel modules are not loaded: [ip_vs ip_vs_rr ip_vs_wrr ip_vs_sh] or no builtin kernel ipvs support: map[ip_vs_wrr:{} ip_vs_sh:{} nf_conntrack_ipv4:{} ip_vs:{} ip_vs_rr:{}]
you can solve this problem with following methods:
 1. Run 'modprobe -- ' to load missing kernel modules;
2. Provide the missing builtin kernel ipvs support

[discovery] Trying to connect to API Server "192.168.0.8:6443"
[discovery] Created cluster-info discovery client, requesting info from "https://192.168.0.8:6443"
[discovery] Requesting info from "https://192.168.0.8:6443" again to validate TLS against the pinned public key
[discovery] Cluster info signature and contents are valid and TLS certificate validates against pinned roots, will use API Server "192.168.0.8:6443"
[discovery] Successfully established connection with API Server "192.168.0.8:6443"
[kubelet] Downloading configuration for the kubelet from the "kubelet-config-1.12" ConfigMap in the kube-system namespace
[kubelet] Writing kubelet configuration to file "/var/lib/kubelet/config.yaml"
[kubelet] Writing kubelet environment file with flags to file "/var/lib/kubelet/kubeadm-flags.env"
[preflight] Activating the kubelet service
[tlsbootstrap] Waiting for the kubelet to perform the TLS Bootstrap...
[patchnode] Uploading the CRI Socket information "/var/run/dockershim.sock" to the Node API object "ubuntu2" as an annotation

This node has joined the cluster:
* Certificate signing request was sent to apiserver and a response was received.
* The Kubelet was informed of the new secure connection details.

Run 'kubectl get nodes' on the master to see this node join the cluster.

等待一會，我們可以在Master節點上使用kubectl get nodes命令來查看節點的狀態：

kubectl get nodes

# 輸出：
# NAME      STATUS   ROLES    AGE     VERSION
# ubuntu1   Ready    master   9h      v1.12.2
# ubuntu2   Ready    <none>   2m24s   v1.12.2

如上全部Ready，大功告成，我們可以運行一些命令來測試一下。

測試

首先驗證kube-apiserver, kube-controller-manager, kube-scheduler, pod network 是否正常：

# 部署一個 Nginx Deployment，包含兩個Pod
# https://kubernetes.io/docs/concepts/workloads/controllers/deployment/
kubectl create deployment nginx --image=nginx:alpine
kubectl scale deployment nginx --replicas=2

# 驗證Nginx Pod是否正確運行，並且會分配10.244.開頭的集群IP
kubectl get pods -l app=nginx -o wide

# 輸出如下：
# NAME                     READY   STATUS    RESTARTS   AGE   IP           NODE      NOMINATED NODE
# nginx-65d5c4f7cc-7pzgp   1/1     Running   0          88s   10.244.1.2   ubuntu2   <none>
# nginx-65d5c4f7cc-l2h26   1/1     Running   0          82s   10.244.1.3   ubuntu2   <none>

再驗證一下kube-proxy是否正常：

# 以 NodePort 方式對外提供服務 https://kubernetes.io/docs/concepts/services-networking/connect-applications-service/
kubectl expose deployment nginx --port=80 --type=NodePort

# 查看集群外可訪問的Port
kubectl get services nginx

# 輸出如下：
# NAME    TYPE       CLUSTER-IP       EXTERNAL-IP   PORT(S)        AGE
# nginx   NodePort   10.110.142.125   <none>        80:30092/TCP   7s

# 可以通過任意 NodeIP:Port 在集群外部訪問這個服務，本示例中部署的2台集群IP分別是192.168.0.8和192.168.0.7
curl http://192.168.0.8:30092
curl http://192.168.0.7:30092

最后驗證一下dns, pod network是否正常：

# 運行Busybox並進入交互模式
kubectl run -it curl --image=radial/busyboxplus:curl

# 輸入`nslookup nginx`查看是否可以正確解析出集群內的IP，已驗證DNS是否正常
[ root@curl-5cc7b478b6-tlf46:/ ]$ nslookup nginx

# 輸出如下：
# Server:    10.96.0.10
# Address 1: 10.96.0.10 kube-dns.kube-system.svc.cluster.local
# 
# Name:      nginx
# Address 1: 10.110.142.125 nginx.default.svc.cluster.local

# 通過服務名進行訪問，驗證kube-proxy是否正常
[ root@curl-5cc7b478b6-tlf46:/ ]$ curl http://nginx/

# 輸出如下：
# <!DOCTYPE html> ---省略

# 分別訪問一下2個Pod的內網IP，驗證跨Node的網絡通信是否正常
[ root@curl-5cc7b478b6-tlf46:/ ]$ curl http://10.244.1.2/
[ root@curl-5cc7b478b6-tlf46:/ ]$ curl http://10.244.1.3/

驗證通過，集群搭建成功，接下來我們就可以參考官方文檔來部署其他服務，愉快的玩耍了。

卸載集群

想要撤銷kubeadm執行的操作，首先要排除節點，並確保該節點為空, 然后再將其關閉。

在Master節點上運行：

kubectl drain <node name> --delete-local-data --force --ignore-daemonsets
kubectl delete node <node name>

然后在需要移除的節點上，重置kubeadm的安裝狀態：

sudo kubeadm reset

如果你想重新配置集群，使用新的參數重新運行kubeadm init或者kubeadm join即可。

參考資料

• Creating a single master cluster with kubeadm