Kubernetes 1.4 基础篇：kubeadm方式安装

kubernetes在9月份推出了1.4的版本，在这个版本中最招人眼球的就是它推出了kubeadm的方式进行安装，简化了很多流程。终于，至少这个版本的kubernetes像docker1.12内置的swarm一样方便的创建集群了。本文将会介绍一下如何在centos7下安装单节点可用的集群。亲测可用。



2016/9/26 kubernetes发布了1.4版，自称“Making it easy to run on Kubernetes anywhere”，还增加了一些特性，该版本最引起关注的可能就是kubeadm的引入了。kubernetes1.4的新增特性中第一个被介绍的就是“两条语句创建集群（Cluster creation with two commands）”了。

kubernetes 1.4新特性

两条语句创建集群

Stateful应用支持的扩展

Cluster federation API additions

容器安全支持（Pod Security Policy+Apparmor）

Infrastructure enhancements（swagger2.0的支持等）

Dashboard UI（能完成CLI的90%操作）

…

安装准备

项番	详细
No.1	一台或多台CentOS7
No.2	大于或等于1G的内存
No.3	网络可以连接

Step 1/4 ：安装kubelet和kubeadm

具体步骤参看如下：

# cat <<EOF > /etc/yum.repos.d/kubernetes.repo
[kubernetes]
name=Kubernetes
baseurl=http://yum.kubernetes.io/repos/kubernetes-el7-x86_64
enabled=1
gpgcheck=1
repo_gpgcheck=1
gpgkey=https://packages.cloud.google.com/yum/doc/yum-key.gpg https://packages.cloud.google.com/yum/doc/rpm-package-key.gpg EOF
# setenforce 0
# yum install -y docker kubelet kubeadm kubectl kubernetes-cni
# systemctl enable docker && systemctl start docker
# systemctl enable kubelet && systemctl start kubelet

注意事项：因为https://packages.cloud.google.com/yum/doc/yum-key.gpg，这个可能很多用户无法访问，可以不用check，设定gpgcheck=0，直接执行即可。

# cat <<EOF > /etc/yum.repos.d/kubernetes.repo
[kubernetes]
name=Kubernetes
baseurl=http://yum.kubernetes.io/repos/kubernetes-el7-x86_64
enabled=1
gpgcheck=0
EOF
#

安装参照

[root@liumiaocn ~]# cat <<EOF > /etc/yum.repos.d/kubernetes.repo
> [kubernetes]
> name=Kubernetes
> baseurl=http://yum.kubernetes.io/repos/kubernetes-el7-x86_64
> enabled=1
> gpgcheck=0
> EOF
[root@liumiaocn ~]# setenforce 0
[root@liumiaocn ~]# yum install -y docker kubelet kubeadm kubectl kubernetes-cni 2>&1 >/tmp/k8s_install.log
[root@liumiaocn ~]# systemctl enable docker && systemctl start docker
Created symlink from /etc/systemd/system/multi-user.target.wants/docker.service to /usr/lib/systemd/system/docker.service.
[root@liumiaocn ~]# systemctl enable kubelet && systemctl start kubelet
Created symlink from /etc/systemd/system/multi-user.target.wants/kubelet.service to /etc/systemd/system/kubelet.service.
[root@liumiaocn ~]#

Step 2/4: 初始化Master

命令：kubeadm init

安装参照：提前将google相关的一些image下载到本地会加快构建集群的速度，应该国内的因为网络的原因无法下载，因此将初始化时需要的6个image提前下载到本地。为了方便，可以使用如下脚本直接执行（参数为PREPARE），其将会直接从liumiaocn的dockerhub下直接pull后rename然后untag。

#!/bin/sh

usage(){
echo "Usage: $0 ACTION "
echo "       ACTION: LIST|PREPARE"
}

PROXY="kube-proxy-amd64:v1.4.0"
DISCOVERY="kube-discovery-amd64:1.0"
SCHEDULER="kube-scheduler-amd64:v1.4.0"
CONTROLLER="kube-controller-manager-amd64:v1.4.0"
APISERVER="kube-apiserver-amd64:v1.4.0"
PAUSE="pause-amd64"

ALL_IMAGES="${PROXY} ${DISCOVERY} ${SCHEDULER} ${CONTROLLER} ${APISERVER} ${PAUSE}"

ACTION="$1"

if [ $# -ne 1 ]; then
usage
exit 1
fi

CNT=0
if [ _"PREPARE" = _"${ACTION}" ]; then
echo "Now begin to pull images from liumiaocn"
for image in ${ALL_IMAGES}
do
CNT=`expr ${CNT} + 1`
echo "No.${CNT} : ${image} pull begins ..."
docker pull liumiaocn/${image}
echo "No.${CNT} : ${image} pull ends   ..."
echo "No.${CNT} : ${image} rename      ..."
docker tag liumiaocn/${image}  gcr.io/google_containers/${image}
echo "No.${CNT} : ${image} untag       ..."
docker rmi  liumiaocn/${image}
echo ""
done

echo "All images have been pulled to local as following"

docker images |egrep 'kube|pause'
else
echo "All images of local for k8s as following"
docker images |egrep 'kube|pause'
fi

初始化命令：kubeadm init

[root@liumiaocn ~]# kubeadm init
<master/tokens> generated token: "9a5d26.9223a9e6bb57b85c"
<master/pki> created keys and certificates in "/etc/kubernetes/pki"
<util/kubeconfig> created "/etc/kubernetes/kubelet.conf"
<util/kubeconfig> created "/etc/kubernetes/admin.conf"
<master/apiclient> created API client configuration
<master/apiclient> created API client, waiting for the control plane to become ready
<master/apiclient> all control plane components are healthy after 55.613825 seconds
<master/apiclient> waiting for at least one node to register and become ready
<master/apiclient> first node is ready after 2.251018 seconds
<master/discovery> created essential addon: kube-discovery, waiting for it to become ready
<master/discovery> kube-discovery is ready after 5.219802 seconds
<master/addons> created essential addon: kube-proxy
<master/addons> created essential addon: kube-dns

Kubernetes master initialised successfully!

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

kubeadm join --token 9a5d26.9223a9e6bb57b85c 16.157.246.203
[root@liumiaocn ~]#

[root@liumiaocn ~]#  kubectl taint nodes --all dedicated-
node "liumiaocn" tainted
[root@liumiaocn ~]# kubectl get nodes
NAME        STATUS    AGE
liumiaocn   Ready     5m
[root@liumiaocn ~]#

总结: kubeadm init进行不下去的主要原因.

原因1：CentOS7下kubernetes1.4的安装有个bug需要执行setenforce 0设定selinux才可，据kubernetes的issue说明后续的版本应该不会再有问题。

原因2：事前container没有提前下载下来，因为各种网络问题一直无法继续

原因3：CentOS7下，因为网络设备编号没有eth0而产生的问题，有可能是当前kubernetes的一个bug，毕竟还只是内测程度的版本，要求不要太高。可以查看系统/var/log/messages即可看出问题所在。

Step 3/4: kubeadm join

kubeadm init之后master就ready了，然后使用一个一个地kubeadm join就完成了集群的创建。方式跟今年合体到docker里面的swarm如出一辙。

看来这个是kubernetes对docker的一大反击，docker在模仿kubernetes的容器管理的时候，kubernetes也开始模仿docker swarm创建集群。

命令：kubeadm join --token 9a5d26.9223a9e6bb57b85c 16.157.246.203

Step 4/4: weave设定

[root@liumiaocn ~]# kubectl apply -f https://git.io/weave-kube daemonset "weave-net" created
[root@liumiaocn ~]#

版本确认

[root@liumiaocn ~]# kubectl version
Client Version: version.Info{Major:"1", Minor:"4", GitVersion:"v1.4.0", GitCommit:"a16c0a7f71a6f93c7e0f222d961f4675cd97a46b", GitTreeState:"clean", BuildDate:"2016-09-26T18:16:57Z", GoVersion:"go1.6.3", Compiler:"gc", Platform:"linux/amd64"}
Server Version: version.Info{Major:"1", Minor:"4", GitVersion:"v1.4.0", GitCommit:"a16c0a7f71a6f93c7e0f222d961f4675cd97a46b", GitTreeState:"clean", BuildDate:"2016-09-26T18:10:32Z", GoVersion:"go1.6.3", Compiler:"gc", Platform:"linux/amd64"}
[root@liumiaocn ~]# kubeadm version
kubeadm version: version.Info{Major:"1", Minor:"5+", GitVersion:"v1.5.0-alpha.0.1534+cf7301f16c0363-dirty", GitCommit:"cf7301f16c036363c4fdcb5d4d0c867720214598", GitTreeState:"dirty", BuildDate:"2016-09-27T18:10:39Z", GoVersion:"go1.6.3", Compiler:"gc", Platform:"linux/amd64"}
[root@liumiaocn ~]#

kubeadm确实有其独到之处，比1.4之前的kubernetes-ansible或者minikube相比方便很多。另外1.4版本的UI也有所强化，我们将在后面进行进一步的介绍。