Kubernetes集群搭建

全面解析 Coco AI，一款开源、免费的智能搜索与知识库工具

0. 概要

使用kubeadm搭建一个单节点kubernets实例,仅供学习. 运行环境和软件概要如下:

以上系统和软件基本是2018.11截止最新的状态, 其中docker需要注意必须安装k8s支持到的版本.

1. 安装步骤

• 关闭系交换分区

swapoff -a

• 安装运行时, 默认使用docker, 安装docker即可

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

• 安装kubeadm 一下命令和官网的命令一致, 但是是包源改为阿里云

apt-get update && apt-get install -y apt-transport-https

curl -s https://mirrors.aliyun.com/kubernetes/apt/doc/apt-key.gpg | apt-key add -

cat <<EOF >/etc/apt/sources.list.d/kubernetes.list
deb https://mirrors.aliyun.com/kubernetes/apt/ kubernetes-xenial main
EOF

apt-get update

apt-get install -y kubelet kubeadm kubectl

apt-mark hold kubelet kubeadm kubectl

2. 使用kubeadm创建集群

2.1 准备镜像

因为国内是访问不到k8s.gcr.io所以需要将需要的镜像提前下载, 这次采用从阿里云镜像仓库下载, 并修改下载后的镜像tag为k8s.gcr.io

# a. 查看都需要哪些镜像需要下载
kubeadm config images list --kubernetes-version=v1.12.3

k8s.gcr.io/kube-apiserver:v1.12.3
k8s.gcr.io/kube-controller-manager:v1.12.3
k8s.gcr.io/kube-scheduler:v1.12.3
k8s.gcr.io/kube-proxy:v1.12.3
k8s.gcr.io/pause:3.1
k8s.gcr.io/etcd:3.2.24
k8s.gcr.io/coredns:1.2

# b. 创建一个自动处理脚本下载镜像->重新tag->删除老tag
vim ./load_images.sh

#!/bin/bash
### config the image map
declare -A images map=()
images["k8s.gcr.io/kube-apiserver:v1.12.3"]="registry.cn-hangzhou.aliyuncs.com/google_containers/kube-apiserver:v1.12.3"
images["k8s.gcr.io/kube-controller-manager:v1.12.3"]="registry.cn-hangzhou.aliyuncs.com/google_containers/kube-controller-manager:v1.12.3"
images["k8s.gcr.io/kube-scheduler:v1.12.3"]="registry.cn-hangzhou.aliyuncs.com/google_containers/kube-scheduler:v1.12.3"
images["k8s.gcr.io/kube-proxy:v1.12.3"]="registry.cn-hangzhou.aliyuncs.com/google_containers/kube-proxy:v1.12.3"
images["k8s.gcr.io/pause:3.1"]="registry.cn-hangzhou.aliyuncs.com/google_containers/pause:3.1"
images["k8s.gcr.io/etcd:3.2.24"]="registry.cn-hangzhou.aliyuncs.com/google_containers/etcd:3.2.24"
images["k8s.gcr.io/coredns:1.2.2"]="registry.cn-hangzhou.aliyuncs.com/google_containers/coredns:1.2.2"
### re-tag foreach
for key in ${!images[@]}
do
	docker pull ${images[$key]}
	docker tag ${images[$key]} $key
	docker rmi ${images[$key]}
done
### check
docker images

# c. 执行脚本准镜像
sudo chmod +x load_images.sh
./load_images.sh

2.2 初始化集群(master)

初始化需要指定至少两个参数:

• kubernetes-version: 方式kubeadm访问外网获取版本
• pod-network-cidr: flannel网络插件配置需要

### 执行初始化命令
sudo kubeadm init --kubernetes-version=v1.12.3 --pod-network-cidr=10.244.0.0/16

### 最后的结果如下
... ...
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 192.168.132.152:6443 --token ymny55.4jlbbkxiggmn9ezh --discovery-token-ca-cert-hash sha256:70265fafdb22d524c15616543d0b76527c686329221340b3b8da3652abed46b9

2.3 根据成功信息配置非管理员账号使用kubectl

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

使用非root账号查看节点情况:

kubectl get nodes

NAME           STATUS     ROLES    AGE   VERSION
servermaster   NotReady   master   28m   v1.12.3

发现有一个master节点, 但是状态是NotReady, 这里需要做一个决定:

如果希望是单机则执行如下

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

如果希望搭建继续, 则继续后续步骤, 此时主节点状态可以忽略.

2.4 应用网络插件

查看kube-flannel.yml文件内容, 复制到本地文件避免terminal无法远程获取

kubectl apply -f kube-flannel.yml

clusterrole.rbac.authorization.k8s.io/flannel created
clusterrolebinding.rbac.authorization.k8s.io/flannel created
serviceaccount/flannel created
configmap/kube-flannel-cfg created
daemonset.extensions/kube-flannel-ds-amd64 created
daemonset.extensions/kube-flannel-ds-arm64 created
daemonset.extensions/kube-flannel-ds-arm created
daemonset.extensions/kube-flannel-ds-ppc64le created
daemonset.extensions/kube-flannel-ds-s390x created

2.5 新建worker节点

worker节点新建参考[1. 安装步骤]在另外一台服务器上新建即可, worker节点不用准备2.1~2.3及之后的所有步骤, 仅需完成基本安装, 安装完毕进入新的worker节点, 执行上一步最后得到join命令:

kubeadm join 192.168.132.152:6443 --token ymny55.4jlbbkxiggmn9ezh --discovery-token-ca-cert-hash sha256:70265fafdb22d524c15616543d0b76527c686329221340b3b8da3652abed46b9

... ...
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.

2.6 检查集群(1 master, 1 worker)

kubectl get nodes

NAME           STATUS   ROLES    AGE   VERSION
servermaster   Ready    master   94m   v1.12.3
worker01       Ready    <none>   54m   v1.12.3

2.7 创建dashboard

复制kubernetes-dashboard.yaml内容到本地文件, 方式命令行无法访问远程文件, 编辑最后一个配置Dashboard Service, 增加type和nodePort, 结果如下:

# ------------------- Dashboard Service ------------------- #

kind: Service
apiVersion: v1
metadata:
  labels:
    k8s-app: kubernetes-dashboard
  name: kubernetes-dashboard
  namespace: kube-system
spec:
  ports:
    - port: 443
      targetPort: 8443
      nodePort: 30000
  type: NodePort
  selector:
    k8s-app: kubernetes-dashboard

在master节点执行创建dashboard服务的命令:

kubectl create -f kubernetes-dashboard.yaml 

secret/kubernetes-dashboard-certs created
serviceaccount/kubernetes-dashboard created
role.rbac.authorization.k8s.io/kubernetes-dashboard-minimal created
rolebinding.rbac.authorization.k8s.io/kubernetes-dashboard-minimal created
deployment.apps/kubernetes-dashboard created
service/kubernetes-dashboard created

浏览器输入worker节点ip和端口使用https访问如下:https://my.worker01.local:30000/#!/login 即可以验证dashboard是否安装成功.

2.8 登录Dashboard

通过kubectl获取secret,然后在获取详细的token,复制到上一步中登录页选择Token(令牌), 即可以登录

### 查看秘钥, 列出所有kube-system命名空间下秘钥
kubectl -n kube-system get secret

NAME                                             TYPE                                  DATA   AGE
clusterrole-aggregation-controller-token-vxzmt   kubernetes.io/service-account-token   3      10h

### 查看token, 这里选择clusterrole-aggregation-controller-token-*****的秘钥
kubectl -n kube-system describe secret clusterrole-aggregation-controller-token-vxzmt

Name:         clusterrole-aggregation-controller-token-vxzmt
Namespace:    kube-system
Labels:       <none>
Annotations:  kubernetes.io/service-account.name: clusterrole-aggregation-controller
              kubernetes.io/service-account.uid: dfb9d9c3-f646-11e8-9861-000c29b7e604

Type:  kubernetes.io/service-account-token

Data
====
ca.crt:     1025 bytes
namespace:  11 bytes
token:      eyJhbGciOiJSUzI1NiIsImtpZCI6IiJ9.eyJpc3MiOiJrdWJlcm5ldGVzL3NlcnZpY2VhY2NvdW50Iiwia3ViZXJuZXRlcy5pby9zZXJ2aWNlYWNjb3VudC9uYW1lc3BhY2UiOiJrdWJlLXN5c3RlbSIsImt1YmVybmV0ZXMuaW8vc2VydmljZWFjY291bnQvc2VjcmV0Lm5hbWUiOiJjbHVzdGVycm9sZS1hZ2dyZWdhdGlvbi1jb250cm9sbGVyLXRva2VuLXZ4em10Iiwia3ViZXJuZXRlcy5pby9zZXJ2aWNlYWNjb3VudC9zZXJ2aWNlLWFjY291bnQubmFtZSI6ImNsdXN0ZXJyb2xlLWFnZ3JlZ2F0aW9uLWNvbnRyb2xsZXIiLCJrdWJlcm5ldGVzLmlvL3NlcnZpY2VhY2NvdW50L3NlcnZpY2UtYWNjb3VudC51aWQiOiJkZmI5ZDljMy1mNjQ2LTExZTgtOTg2MS0wMDBjMjliN2U2MDQiLCJzdWIiOiJzeXN0ZW06c2VydmljZWFjY291bnQ6a3ViZS1zeXN0ZW06Y2x1c3RlcnJvbGUtYWdncmVnYXRpb24tY29udHJvbGxlciJ9.MfjiMrmyKl1GUci1ivD5RNrzo_s6wxXwFzgM_3NIAmTCfRQreYdhat3yyd6agaCLpUResnNC0ZGRi4CBs_Jgjqkovhb80V05_YVIvCrlf7xHxBKEtGfkJ-qLDvtAwR5zrXNNd0Ge8hTRxw67gZ3lGMkPpw5nfWmc0rzk90xTTQD1vAtrHMvxjr3dVXph4rT8GNuCSXA_J6o2AwYUbaKCc2ugdx8t8zX6oFJfVcw0ZNYYYIyxoXzzfhdppORtKR9t9v60KsI_-q0TxY-TU-JBtzUJU-hL6lB5MOgoBWpbQiV-aG8Ov74nDC54-DH7EhYEzzsLci6uUQCPlHNvLo_J2A

3. 遇到问题

• master搭建好了, worker也join了get nodes发现还是NotReady状态

  原因: 太复杂说不清楚任然是一个k8s issue, 查看issue基本可以确定是cni(Container Network Interface)问题,而flannel覆盖修改了这个问题

  解决方法: 安装flannel插件(kubectl apply -f kube-flannel.yml)

• 配置错误重新开始搭建集群

  解决方案: kubeadm reset

• 不能访问dashboard

  原因: Back-off pulling image "k8s.gcr.io/kubernetes-dashboard-amd64:v1.10.0

  解决方案:

  1. 修改 kubernetes-dashboard-ce.yaml 文件中的 k8s.gcr.io/kubernetes-dashboard-amd64:v1.10.0 为 registry.cn-hangzhou.aliyuncs.com/google_containers/kubernetes-dashboard-amd64:v1.10.0

  2. 提前下载好镜像并配置好tag, 注意下载的位置worker节点, 可以通过: kubectl describe pod kubernetes-dashboard-85477d54d7-wzt7 -n kube-system 查看比较具体的信息

• 如何增加token失效时间

  原因: 默认15分钟

  解决方法:

  • 如果创建dashboard前: 可以修改kubernetes-dashboard.yaml文件中Dashboard Deployment部分关于containers的args部分添加一行: - --token-ttl=86400 即可,数字自定义单位是秒 如下:

  ... ...
    - name: kubernetes-dashboard
  image: registry.cn-hangzhou.aliyuncs.com/google_containers/kubernetes-dashboard-amd64:v1.10.0
  ports:
  - containerPort: 8443
    protocol: TCP
  args:
    - --auto-generate-certificates
    - --token-ttl=86400
    # Uncomment the following line to manually specify Kubernetes API server Host
    # If not specified, Dashboard will attempt to auto discover the API server and connect
    # to it. Uncomment only if the default does not work.
    # - --apiserver-host=http://my-address:port
  volumeMounts:
  ... ...
  
  

  • 如果创建dashboard后: kubectl -n kube-system edit deployment kubernetes-dashboard, 在Deployment部分关于containers的args部分添加一行: - --token-ttl=86400 即可, 和创建前修改一样的方式一样

4. 效率&技巧

4.1 kubeadm 自动完成

使用 kubeadm completion --help 查看使用详情,这里直接贴出bash的自动完成命令

kubeadm completion bash > ~/.kube/kubeadm_completion.bash.inc

printf "\n# Kubeadm shell completion\nsource '$HOME/.kube/kubeadm_completion.bash.inc'\n" >> $HOME/.bash_profile

source $HOME/.bash_profile

4.2 kubectl 自动完成

使用 kubectl completion --help 查看使用详情,这里直接贴出bash的自动完成命令, 注意第二行命令不要一次性复制,先复制第一行printf再复制剩余.

kubectl completion bash > ~/.kube/completion.bash.inc

printf "
# Kubectl shell completion
source '$HOME/.kube/completion.bash.inc'
" >> $HOME/.bash_profile

source $HOME/.bash_profile

4.3 使用私有 docker registry

创建 secret, 然后增加添加 imagePullSecrets 配置在指定image的地方. 创建和查看secret如下:

kubectl create secret docker-registry regcred --docker-server=registry.domain.cn:5001 --docker-username=xxxxx --docker-password=xxxxx --docker-email=jimmy.w@aliyun.com

kubectl get secret regcred --output=yaml

kubectl get secret regcred --output="jsonpath={.data.\.dockerconfigjson}" | base64 --decode

配置 imagePullSecrets 如下:

... ...
containers:
- name: mirage
  image: registry.domain.cn:5001/mirage:latest
imagePullSecrets:
  - name:  regcred
ports:
- containerPort: 3008
  protocol: TCP
volumeMounts:
... ...

4.4 使用 HostAliases 向 Pod /etc/hosts 文件添加条目

如果有一些特别的入口或者以前放置到/etc/hosts中的可以通过配置hostAliases进行配置, 作用和本地的hosts一样, 且这些hostAlieas配置会放置到容器/etc/hosts中, 具体使用如下:

apiVersion: v1
kind: Pod
metadata:
  name: hostaliases-pod
spec:
  hostAliases:
  - ip: "127.0.0.1"
    hostnames:
    - "foo.local"
    - "bar.local"
  - ip: "10.1.2.3"
    hostnames:
    - "foo.remote"
    - "bar.remote"
  containers:
  - name: cat-hosts
    image: busybox
    command:
    - cat
    args:
    - "/etc/hosts"

5. 参考资料

安装kuadmin相关:

• Installing kubeadm
• kubernetes-for-china

创建集群相关:

• Creating a single master cluster with kubeadm
• 基于Ubuntu16.04和kubeadm部署kuernetes