之前测试了离线环境下使用二进制方法安装配置Kubernetes集群的方法，安装的过程中听说 kubeadm 安装配置集群更加方便，因此试着折腾了一下。安装过程中，也有一些坑，相对来说操作上要比二进制方便一点，毕竟不用手工创建那么多的配置文件，但是对于了解Kubernetes的运作方式，可能不如二进制方式好。同时，因为kubeadm方式，很多集群依赖的组件都是以容器方式运行在Master节点上，感觉对于虚拟机资源的消耗要比二进制方式厉害。
0. kubeadm 介绍与准备工作
kubeadm is designed to be a simple way for new users to start trying Kubernetes out, possibly for the first time, a way for existing users to test their application on and stitch together a cluster easily, and also to be a building block in other ecosystem and/or installer tool with a larger scope.
kubeadm是一个python写的项目，代码在这里，用来帮助快速部署Kubernetes集群环境，但是目前仅仅是作为测试环境使用，如果你想在生产环境使用，可是要三思。
本文所用的环境：
- 虚拟机软件：VirtualBox
- 操作系统：Centos 7.3 minimal 安装
- 网卡：两块网卡，一块 Host-Only方式，一块 Nat 方式。
- 网络规划：
- Master:192.168.0.101
- Node:192.168.0.102-104
0.1 关掉 selinux
$ setenforce 0
$ sed -i "s/^SELINUX=enforcing/SELINUX=disabled/g" /etc/sysconfig/selinux
0.2 关掉防火墙
$ systemctl stop firewalld
$ systemctl disable firewalld
0.3 关闭 swap
$ swapoff -a
$ sed -i 's/.*swap.*/#&/' /etc/fstab
0.4 配置转发参数
$ cat <<EOF > /etc/sysctl.d/k8s.conf
net.bridge.bridge-nf-call-ip6tables = 1
net.bridge.bridge-nf-call-iptables = 1
EOF
$ sysctl --system
0.5 设置国内 yum 源
$ cat <<EOF > /etc/yum.repos.d/kubernetes.repo
[kubernetes]
name=Kubernetes
baseurl=https://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64/
enabled=1
gpgcheck=1
repo_gpgcheck=1
gpgkey=https://mirrors.aliyun.com/kubernetes/yum/doc/yum-key.gpg https://mirrors.aliyun.com/kubernetes/yum/doc/rpm-package-key.gpg
EOF
0.6 安装一些必备的工具
$ yum install -y epel-release
$ yum install -y net-tools wget vim ntpdate
1. 安装 kubeadm 必须的软件，在所有节点上运行
1.1 安装Docker
$ yum install -y docker
$ systemctl enable docker && systemctl start docker
$ #设置系统服务，如果不设置后面 kubeadm init 的时候会有 warning
$ systemctl enable docker.service
如果想要用二进制方法安装最新版本的Docker，可以参考我之前的文章在Redhat 7.3中采用离线方式安装Docker
1.2 安装kubeadm、kubectl、kubelet
$ yum install -y kubelet kubeadm kubectl kubernetes-cni
$ systemctl enable kubelet && systemctl start kubelet
这一步之后kubelet还不能正常运行，还处于下面的状态。
The kubelet is now restarting every few seconds, as it waits in a crashloop for kubeadm to tell it what to do.
2. 安装Master节点
因为国内没办法访问Google的镜像源，变通的方法是从其他镜像源下载后，修改tag。执行下面这个Shell脚本即可。
#!/bin/bash
images=(kube-proxy-amd64:v1.11.0 kube-scheduler-amd64:v1.11.0 kube-controller-manager-amd64:v1.11.0 kube-apiserver-amd64:v1.11.0
etcd-amd64:3.2.18 coredns:1.1.3 pause-amd64:3.1 kubernetes-dashboard-amd64:v1.8.3 k8s-dns-sidecar-amd64:1.14.9 k8s-dns-kube-dns-amd64:1.14.9
k8s-dns-dnsmasq-nanny-amd64:1.14.9 )
for imageName in ${images[@]} ; do
docker pull registry.cn-hangzhou.aliyuncs.com/k8sth/$imageName
docker tag registry.cn-hangzhou.aliyuncs.com/k8sth/$imageName k8s.gcr.io/$imageName
#docker rmi registry.cn-hangzhou.aliyuncs.com/k8sth/$imageName
done
docker tag da86e6ba6ca1 k8s.gcr.io/pause:3.1
接下来执行Master节点的初始化，因为我的虚拟机是双网卡，需要指定apiserver的监听地址。
[root@devops-101 ~]# kubeadm init --kubernetes-version=v1.11.0 --pod-network-cidr=10.244.0.0/16 --apiserver-advertise-address=192.168.0.101
[init] using Kubernetes version: v1.11.0
[preflight] running pre-flight checks
I0724 08:36:35.636931 3409 kernel_validator.go:81] Validating kernel version
I0724 08:36:35.637052 3409 kernel_validator.go:96] Validating kernel config
[WARNING Hostname]: hostname "devops-101" could not be reached
[WARNING Hostname]: hostname "devops-101" lookup devops-101 on 172.20.10.1:53: no such host
[WARNING Service-Kubelet]: kubelet service is not enabled, please run 'systemctl enable kubelet.service'
[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 [devops-101 kubernetes kubernetes.default kubernetes.default.svc kubernetes.default.svc.cluster.local] and IPs [10.96.0.1 192.168.0.101]
[certificates] Generated apiserver-kubelet-client certificate and key.
[certificates] Generated sa key and public 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 [devops-101 localhost] and IPs [127.0.0.1 ::1]
[certificates] Generated etcd/peer certificate and key.
[certificates] etcd/peer serving cert is signed for DNS names [devops-101 localhost] and IPs [192.168.0.101 127.0.0.1 ::1]
[certificates] Generated etcd/healthcheck-client certificate and key.
[certificates] Generated apiserver-etcd-client certificate and key.
[certificates] valid certificates and keys now exist in "/etc/kubernetes/pki"
[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 46.002877 seconds
[uploadconfig] storing the configuration used in ConfigMap "kubeadm-config" in the "kube-system" Namespace
[kubelet] Creating a ConfigMap "kubelet-config-1.11" in namespace kube-system with the configuration for the kubelets in the cluster
[markmaster] Marking the node devops-101 as master by adding the label "node-role.kubernetes.io/master=''"
[markmaster] Marking the node devops-101 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 "devops-101" as an annotation
[bootstraptoken] using token: wkj0bo.pzibll6rd9gyi5z8
[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 192.168.0.101:6443 --token wkj0bo.pzibll6rd9gyi5z8 --discovery-token-ca-cert-hash sha256:51985223a369a1f8c226f3ccdcf97f4ad5ff201a7c8c708e1636eea0739c0f05
看到以上信息表示Master节点已经初始化成功了。如果需要用普通用户管理集群，可以按照提示进行操作，如果是使用root用户管理，执行下面的命令。
[root@devops-101 ~]# export KUBECONFIG=/etc/kubernetes/admin.conf
[root@devops-101 ~]# kubectl get nodes
NAME STATUS ROLES AGE VERSION
devops-101 NotReady master 7m v1.11.1
[root@devops-101 ~]# kubectl get pods --all-namespaces
NAMESPACE NAME READY STATUS RESTARTS AGE
kube-system coredns-78fcdf6894-8sd6g 0/1 Pending 0 7m
kube-system coredns-78fcdf6894-lgvd9 0/1 Pending 0 7m
kube-system etcd-devops-101 1/1 Running 0 6m
kube-system kube-apiserver-devops-101 1/1 Running 0 6m
kube-system kube-controller-manager-devops-101 1/1 Running 0 6m
kube-system kube-proxy-bhmj8 1/1 Running 0 7m
kube-system kube-scheduler-devops-101 1/1 Running 0 6m
可以看到节点还没有Ready，dns的两个pod也没不正常，还需要安装网络配置。
3. Master节点的网络配置
这里我选用了 Flannel 的方案。
kubeadm only supports Container Network Interface (CNI) based networks (and does not support kubenet).
修改系统设置，创建 flannel 网络。
[root@devops-101 ~]# sysctl net.bridge.bridge-nf-call-iptables=1
net.bridge.bridge-nf-call-iptables = 1
[root@devops-101 ~]# kubectl apply -f https://raw.githubusercontent.com/coreos/flannel/v0.10.0/Documentation/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 created
flannel 默认会使用主机的第一张网卡，如果你有多张网卡，需要通过配置单独指定。修改 kube-flannel.yml 中的以下部分
containers:
- name: kube-flannel
image: quay.io/coreos/flannel:v0.10.0-amd64
command:
- /opt/bin/flanneld
args:
- --ip-masq
- --kube-subnet-mgr
- --iface=enp0s3 #指定内网网卡
执行成功后，Master并不能马上变成Ready状态，稍等几分钟，就可以看到所有状态都正常了。
[root@devops-101 ~]# kubectl get pods --all-namespaces
NAMESPACE NAME READY STATUS RESTARTS AGE
kube-system coredns-78fcdf6894-8sd6g 1/1 Running 0 14m
kube-system coredns-78fcdf6894-lgvd9 1/1 Running 0 14m
kube-system etcd-devops-101 1/1 Running 0 13m
kube-system kube-apiserver-devops-101 1/1 Running 0 13m
kube-system kube-controller-manager-devops-101 1/1 Running 0 13m
kube-system kube-flannel-ds-6zljr 1/1 Running 0 48s
kube-system kube-proxy-bhmj8 1/1 Running 0 14m
kube-system kube-scheduler-devops-101 1/1 Running 0 13m
[root@devops-101 ~]# kubectl get nodes
NAME STATUS ROLES AGE VERSION
devops-101 Ready master 14m v1.11.1
4. 加入节点
Node节点的加入集群前，首先需要按照本文的第0节和第1节做好准备工作，然后下载镜像。
$ docker pull registry.cn-hangzhou.aliyuncs.com/k8sth/kube-proxy-amd64:v1.11.0
$ docker pull registry.cn-hangzhou.aliyuncs.com/k8sth/pause-amd64:3.1
$ docker tag registry.cn-hangzhou.aliyuncs.com/k8sth/pause-amd64:3.1 k8s.gcr.io/pause-amd64:3.1
$ docker tag registry.cn-hangzhou.aliyuncs.com/k8sth/kube-proxy-amd64:v1.11.0 k8s.gcr.io/kube-proxy-amd64:v1.11.0
$ docker tag registry.cn-hangzhou.aliyuncs.com/k8sth/pause-amd64:3.1 k8s.gcr.io/pause:3.1
最后再根据Master节点的提示加入集群。
$ kubeadm join 192.168.0.101:6443 --token wkj0bo.pzibll6rd9gyi5z8 --discovery-token-ca-cert-hash sha256:51985223a369a1f8c226f3ccdcf97f4ad5ff201a7c8c708e1636eea0739c0f05
节点的启动也需要一点时间，稍后再到Master上查看状态。
[root@devops-101 ~]# kubectl get nodes
NAME STATUS ROLES AGE VERSION
devops-101 Ready master 1h v1.11.1
devops-102 Ready <none> 11m v1.11.1
我把安装中需要用到的一些命令整理成了几个脚本，放在我的Github上，大家可以下载使用。
X. 坑
pause:3.1
安装的过程中，发现kubeadmin会找 pause:3.1 的镜像，所以需要重新 tag 。
$ docker tag registry.cn-hangzhou.aliyuncs.com/k8sth/pause-amd64:3.1 k8s.gcr.io/pause:3.1
两台服务器时间不同步。
报错信息
[discovery] Failed to request cluster info, will try again: [Get https://192.168.0.101:6443/api/v1/namespaces/kube-public/configmaps/cluster-info: x509: certificate has expired or is not yet valid]
解决方法，设定一个时间服务器同步两台服务器的时间。
$ ntpdate ntp1.aliyun.com
参考资料
- centos7.3 kubernetes/k8s 1.10 离线安装
- Kubeadm安装Kubernetes环境
- Steps to install kubernetes
- kubeadm reference guide
- kubeadm安装Kubernetes V1.10集群详细文档
- kubeadm reference
- kubeadm搭建kubernetes1.7.5集群
- 安装部署 Kubernetes 集群
- linux 命令 ---- 同步当前服务器时间
- CentOS 7.4 安装 K8S v1.11.0 集群所遇到的问题
- 使用kubeadm部署kubernetes
