6x's notepad

0x5f3759df

0%

从零开始搭建k8s集群环境(三)——搭建Master节点

发表于 更新于 分类于 阅读次数： 4543

从零开始搭建 k8s 集群环境系列笔记：

经过一夜的休整,整个人又活力满满呢哈哈！
那么，今天就正式开始搭建 k8s 集群环境。先从 Master 节点开始，了解的朋友应该知道，Master 节点搭建因为要下载 Google 的镜像，对于这么这种只有小水管的用户，压力很大啊！不过今天演示我们不使用小水管，ok，看我骚操作

环境说明：

  • Master:VBox虚拟机/Centos_7.5-1804/192.168.56.100
  • Images:VBox虚拟机/Centos_7.5-1804/192.168.56.101
  • Node01:VBox虚拟机/Centos_7.5-1804/192.168.56.102

软件环境说明:

  • docker-ce-17.03.2
  • kubernetes-v1.9.0
  • harbor-v1.4.0

这里依旧使用 Master 机器（这不废话吗）。

一、环境准备

1.关闭交换分区

因为 SWAP 的存在影响 k8s 性能,所以 k8s 默认需要关闭.当然为了一些节省资源的场景，比如运行容器数量较多，可添加 kubelet 参数 --fail-swap-on=false 来解决。

1.暂时关闭SWAP，重启后恢复

swapoff   -a

2.永久关闭SWAP

vim /etc/fstab

# 注释掉SWAP分区项，即可
# swap was on /dev/sda11 during installation
# UUID=0a55fdb5-a9d8-4215-80f7-f42f75644f69 none  swap    sw      0       0

2.Docker & Docker仓库镜像

k8s 组件都是运行在 Docker 容器中的，肯定是需要 Docker 的。
Docker 使用离线版 docker-ce-17.03.2.ce-1.el7.centos.x86_64.rpm
下载地址: https://download.docker.com/linux/centos/7/x86_64/stable/Packages/

wget https://download.docker.com/linux/centos/7/x86_64/stable/Packages/docker-ce-17.03.2.ce-1.el7.centos.x86_64.rpm
wget https://download.docker.com/linux/centos/7/x86_64/stable/Packages/docker-ce-selinux-17.03.2.ce-1.el7.centos.noarch.rpm

# 将安装 Docker 和依赖包
yum install -y docker-ce-*.rpm

# 开机启动
[root@localhost ~]# systemctl enable docker
Created symlink from /etc/systemd/system/multi-user.target.wants/docker.service to /usr/lib/systemd/system/docker.service.

# 启动 docker 服务
[root@localhost ~]# systemctl restart docker

启动起来再看下版本：

# 查看 docker 版本
[root@localhost ~]# docker version
Client:
 Version:      17.03.2-ce
 API version:  1.27
 Go version:   go1.7.5
 Git commit:   f5ec1e2
 Built:        Tue Jun 27 02:21:36 2017
 OS/Arch:      linux/amd64

Server:
 Version:      17.03.2-ce
 API version:  1.27 (minimum version 1.12)
 Go version:   go1.7.5
 Git commit:   f5ec1e2
 Built:        Tue Jun 27 02:21:36 2017
 OS/Arch:      linux/amd64
 Experimental: false

然后就是镜像，虽然k8s需要的组件是从 gcr.io 服务器拉取的，但是集群创建好之后需要的的镜像并不一定~ 所以就按需配置吧~
这里推荐阿里云的，点击查看点击查看点击查看

3.kubeadm、kubectl、kubelet、kubernetes-cni

昨晚辛苦辛苦搞的就是这些。在上一篇文章中已经编译且安装了，这里就略过了。
看下版本：

[root@localhost ~]# kubeadm version
kubeadm version: &version.Info{Major:"1", Minor:"9", GitVersion:"v1.9.0", GitCommit:"925c127ec6b946659ad0fd596fa959be43f0cc05", GitTreeState:"clean", BuildDate:"2017-12-15T20:55:30Z", GoVersion:"go1.9.2", Compiler:"gc", Platform:"linux/amd64"}

[root@localhost ~]# kubectl version
Client Version: version.Info{Major:"1", Minor:"9", GitVersion:"v1.9.0", GitCommit:"925c127ec6b946659ad0fd596fa959be43f0cc05", GitTreeState:"clean", BuildDate:"2017-12-15T21:07:38Z", GoVersion:"go1.9.2", Compiler:"gc", Platform:"linux/amd64"}
The connection to the server localhost:8080 was refused - did you specify the right host or port?
# 额本机的k8s的api服务还没启动,只显示了客户端的信息

另外,如果不自己去编译也可以使用其他 yum 源来安装，比如阿里云里面的版本就很全...

[root@localhost ~]# cat <<EOF > /etc/yum.repos.d/kubernetes.repo
[kubernetes]
name=Kubernetes
baseurl=http://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64
enabled=1
gpgcheck=0
repo_gpgcheck=0
gpgkey=http://mirrors.aliyun.com/kubernetes/yum/doc/yum-key.gpg
       http://mirrors.aliyun.com/kubernetes/yum/doc/rpm-package-key.gpg
EOF

# 更新缓存
[root@localhost ~]# yum makecache

# 使用yum安装
[root@localhost ~]# yum install -y kubeadm kubectl kubelet

启动 kubelet 服务，并设置开机启动

[root@localhost ~]# systemctl restart kubelet
[root@localhost ~]# systemctl enable kubelet
Created symlink from /etc/systemd/system/multi-user.target.wants/kubelet.service to /etc/systemd/system/kubelet.service.

4.关闭防火墙

实际操作发现开着防火墙会出一系列问题,所以还是关掉!

service iptables stop
service firewalld stop

# 取消开机启动
systemctl disable iptables
systemctl disable firewalld

二、加载k8s组件镜像

  • 注意：此步骤非常关键，kubenetes 初始化启动会依赖这些镜像，天朝的网络肯定是拉不下来 google 的镜像的，一般人过了上一关，这一关未必过的去，一定要提前把镜像下载到本地，kubeadm 安装才会继续，下面我会列出来 master 节点和 node 依赖的镜像列表。（备注：考虑到随着 kubernetes 版本一直更新，镜像也可能会有变化，大家可以先执行 kubeadm init 生成配置文件，日志输出到 [apiclient] Created API client, waiting for the control plane to become ready 这一行就会卡住不动了，你可以直接执行 ctrl + c 中止命令执行，然后查看 ls -ltr /etc/kubernetes/manifests/ 中 yaml文件列表，每个文件都会写着镜像的地址和版本）
    在这里我提一个可以解决下载 google 镜像的方法，就是买一台可以下载的机器，安装代理软件，在需要下载 google 镜像的机器的 docker 设置 HTTP_PROXY 配置项，配置好自己的服务代理即可（也可以直接买可以访问到 google 的服务器安装).

1.找到镜像名称

先执行一次初始化命令

[root@localhost ~]# kubeadm init
[init] Using Kubernetes version: v1.9.8
[init] Using Authorization modes: [Node RBAC]
[preflight] Running pre-flight checks.
	[WARNING Hostname]: hostname "localhost.master" could not be reached
	[WARNING Hostname]: hostname "localhost.master" lookup localhost.master on 8.8.8.8:53: no such host
	[WARNING Firewalld]: firewalld is active, please ensure ports [6443 10250] are open or your cluster may not function correctly
	[WARNING FileExisting-crictl]: crictl not found in system path
[certificates] Generated ca certificate and key.
[certificates] Generated apiserver certificate and key.
[certificates] apiserver serving cert is signed for DNS names [localhost.master kubernetes kubernetes.default kubernetes.default.svc kubernetes.default.svc.cluster.local] and IPs [10.96.0.1 10.0.2.15]
[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] Valid certificates and keys now exist in "/etc/kubernetes/pki"
[kubeconfig] Wrote KubeConfig file to disk: "admin.conf"
[kubeconfig] Wrote KubeConfig file to disk: "kubelet.conf"
[kubeconfig] Wrote KubeConfig file to disk: "controller-manager.conf"
[kubeconfig] Wrote KubeConfig file to disk: "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.

成功的卡在了 pull 镜像的位置，我们 Ctrl + C 结束命令，查看一下文件列表：

[root@localhost ~]# ll /etc/kubernetes/manifests/
总用量 16
-rwx------. 1 root root  876 5月  20 12:23 etcd.yaml
-rwx------. 1 root root 2545 5月  20 12:23 kube-apiserver.yaml
-rwx------. 1 root root 2002 5月  20 12:23 kube-controller-manager.yaml
-rwx------. 1 root root  991 5月  20 12:23 kube-scheduler.yaml

找出需要的镜像:

[root@localhost ~]# cat /etc/kubernetes/manifests/* |grep image
    image: gcr.io/google_containers/etcd-amd64:3.1.10
    image: gcr.io/google_containers/kube-apiserver-amd64:v1.9.8
    image: gcr.io/google_containers/kube-controller-manager-amd64:v1.9.8
    image: gcr.io/google_containers/kube-scheduler-amd64:v1.9.8

除了上面四个之外,还需要一个(别问我怎么知道了,说多了都是泪)

gcr.io/google_containers/pause-amd64:3.0
gcr.io/google_containers/kube-proxy-amd64:v1.9.8

2.拿到镜像

由于 gcr.io 实际上是 Google 公司的域名,所以如果要在国内拿到 k8s 镜像下载，主要有四个方法:

  • 1.科学上网
  • 2.找国内现成的镜像
  • 3.使用第三方自动构建服务

科学上网

如果你有一个可以科学上网的 VPN/SS 之类的账户，那么你只要对系统配置网络，让系统能访问上 gcr.io 就可以了，很简单，掏钱就行（充钱能让你的服务变快~！）。
或者你有一台可以访问 gcr.io 的服务器，那么你也可以用它装一个 Docker 然后 pull 下你需要的镜像，然后通过 save 打包传回国内再 load 进目标机器。

docker pull gcr.io/google_containers/pause-amd64:3.0
docker pull gcr.io/google_containers/etcd-amd64:3.1.10
docker pull gcr.io/google_containers/kube-proxy-amd64:v1.9.8
docker pull gcr.io/google_containers/kube-apiserver-amd64:v1.9.8
docker pull gcr.io/google_containers/kube-scheduler-amd64:v1.9.8
docker pull gcr.io/google_containers/kube-controller-manager-amd64:v1.9.8

docker save {id} -o file.tar {tag}
docker load <file.tar

找国内现成的镜像

这个方法，怎么说呢，看运气了如果你使用某些博客提供的全套方案和二进制，那么也许可行。反正我是不会这么做的，知其然还要知其所以然对吧

使用第三方自动构建服务

这个方案原理是利用能够访问外网的第三方服务根据我们的 Dockerfile 构建出镜像，然后我们去下载构建出来的镜像,因为这个镜像 FROM 自墙外的镜像,就变成我们间接下载了我们真正需要的镜像。
当然这个方案网上也有讲，但是用的是 Docker Hub 的镜像构建服务，今天试了试，网速表现一般~ 有兴趣的可以看看：传送门在此，感谢前辈的辛勤付出，为后来者提供了莫大的帮助！

所以（重点来了~赶紧记小抄！），这里以我今天发现的阿里云的自动构建服务特别好用，这里就用它来操作：

PS:目前阿里云容器镜像服务还处于试商用阶段，不仅免费而且免费！最重要的速度快啊！(补加一个:还能自动检查 Dockerfile 变化自动构建新版 讲真，我是不是应该向阿里云申请一笔广告费了~！）

首先你需要一些包含 FROM 上面提到镜像的 Dockerfile，这些文件不用太复杂，只需要 FROM 就可以了,例如:

FROM gcr.io/google_containers/kube-apiserver-amd64:v1.9.8
MAINTAINER lx0758 lx0758@qq.com

当然，你也可以 frok 我的仓库，里面已经包含了需要 Dockerfile 文件：
https://github.com/lx0758/automated-build
目录结构：

automated-build> tree /f
.
│  Dockerfile
│  README.md
└─gcr.io
    └─google_containers
        ├─etcd-amd64
        │      Dockerfile
        ├─kube-apiserver-amd64
        │      Dockerfile
        ├─kube-controller-manager-amd64
        │      Dockerfile
        ├─kube-scheduler-amd64
        │      Dockerfile
        ├─kube-proxy-amd64
        │      Dockerfile
        └─pause-amd64
               Dockerfile

然后去 阿里云容器服务控制台 自行注册&登录，然后创建对应个数的仓库。直到构建出相应的镜像：

创建仓库：
alt

alt

alt

构建镜像:
alt

使用说明:
alt

下拉镜像:

[root@localhost ~]# docker pull registry.cn-shenzhen.aliyuncs.com/lx0758/kube-apiserver-amd64:v1.9.8
Trying to pull repository registry.cn-shenzhen.aliyuncs.com/lx0758/kube-apiserver-amd64 ... 
v1.9.8: Pulling from registry.cn-shenzhen.aliyuncs.com/lx0758/kube-apiserver-amd64
f70adabe43c0: Pull complete 
11f418198626: Pull complete 
Digest: sha256:e0904e02e874cc8d85ad3a92dbbea35eac317938aac79ff2bc2d8e163d97a557
Status: Downloaded newer image for registry.cn-shenzhen.aliyuncs.com/lx0758/kube-apiserver-amd64:v1.9.8

更改镜像 tag:

[root@localhost ~]# docker tag registry.cn-shenzhen.aliyuncs.com/lx0758/kube-apiserver-amd64:v1.9.8 gcr.io/google_containers/kube-apiserver-amd64:v1.9.8
[root@localhost ~]# docker images
REPOSITORY                                                      TAG                 IMAGE ID            CREATED             SIZE
gcr.io/google_containers/kube-apiserver-amd64                   v1.9.8              9df87dd51f23        12 minutes ago      212 MB
registry.cn-shenzhen.aliyuncs.com/lx0758/kube-apiserver-amd64   v1.9.8              9df87dd51f23        12 minutes ago      212 MB

这样,就成功 pull 下来一个需要的镜像,剩余几个大家重复上面的操作就可以了。

因为我在阿里云创建这几个仓库都是公开的,所以大家也许可以直接下拉:

#!/usr/bin/env bash
images=(
pause-amd64:3.0
etcd-amd64:3.1.10
kube-proxy-amd64:v1.9.8
kube-apiserver-amd64:v1.9.8
kube-scheduler-amd64:v1.9.8
kube-controller-manager-amd64:v1.9.8
)
for imageName in ${images[@]} ; do
    docker pull registry.cn-shenzhen.aliyuncs.com/lx0758/$imageName
    docker tag registry.cn-shenzhen.aliyuncs.com/lx0758/$imageName gcr.io/google_containers/$imageName
    docker rmi registry.cn-shenzhen.aliyuncs.com/lx0758/$imageName
done

OK,贴一下我目前的镜像：

[root@localhost ~]# docker images
REPOSITORY                                               TAG                 IMAGE ID            CREATED             SIZE
gcr.io/google_containers/pause-amd64                     3.0                 ec38a2020e09        4 hours ago         747 kB
gcr.io/google_containers/kube-apiserver-amd64            v1.9.8              33c677c83acc        4 hours ago         212 MB
gcr.io/google_containers/etcd-amd64                      3.1.10              1170a52f253a        4 hours ago         193 MB
gcr.io/google_containers/kube-proxy-amd64                v1.9.8              46e3fe19734f        4 hours ago         109 MB
gcr.io/google_containers/kube-scheduler-amd64            v1.9.8              ad84db7950b0        4 hours ago         62.9 MB
gcr.io/google_containers/kube-controller-manager-amd64   v1.9.8              5341036204e5        4 hours ago         139 MB

可以看到,pull 下来的镜像和修改 tag 的镜像各一个。到这里，我们需要的镜像就准备完毕了！

三、初始化 Master 节点

Kubelet 启动确认

为了保险，咱们启动一下

systemctl enable kubelet
systemctl restart kubelet

重置初始化

因为上个步骤我们已经调用了 kubeadm init ，所以直接再次调用的话会报错。需要先重置：

kubeadm reset

初始化

kubeadm init --kubernetes-version=1.9.8 --apiserver-advertise-address=192.168.56.100 --apiserver-cert-extra-sans=10.0.2.15,192.168.56.100 --pod-network-cidr=10.244.0.0/16

加入 --kubernetes-version=1.9.8 是因为 kubeadm 初始化会去查询最新的 k8s 版本，但这个地址在墙外，楼主这里经常网络超时，就直接设置而不再查询。

加入 --apiserver-advertise-address=192.168.56.100 是为了注册这个地址为 ApiService 绑定地址,然后我这个地址是虚拟机之间的通讯地址，这样 Node 才能通过这个地址 join 进来。

加入 --apiserver-cert-extra-sans=10.0.2.15,192.168.56.100 是因为我虚拟机有多张网卡，k8s 生成的鉴权证书时需要包含这些地址，才不会导致使用 kubectl（10.0.2.15） 和 Node join（192.168.56.100） 时出现 x509 错误。

因为我后面安装flannel网络，所以提前加上：--pod-network-cidr=10.244.0.0/16。

激动人心的时刻到来了（真的，你不知道楼主这几天经历了些什么。。。）：

[root@localhost ~]# kubeadm init --kubernetes-version=1.9.8 --apiserver-advertise-address=192.168.56.100 --apiserver-cert-extra-sans=10.0.2.15,192.168.56.100
[init] Using Kubernetes version: v1.9.8
[init] Using Authorization modes: [Node RBAC]
[preflight] Running pre-flight checks.
	[WARNING Hostname]: hostname "localhost.master" could not be reached
	[WARNING Hostname]: hostname "localhost.master" lookup localhost.master on 8.8.8.8:53: no such host
	[WARNING FileExisting-crictl]: crictl not found in system path
[preflight] Starting the kubelet service
[certificates] Generated ca certificate and key.
[certificates] Generated apiserver certificate and key.
[certificates] apiserver serving cert is signed for DNS names [localhost.master kubernetes kubernetes.default kubernetes.default.svc kubernetes.default.svc.cluster.local] and IPs [10.96.0.1 192.168.56.100 10.0.2.15 192.168.56.100]
[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] Valid certificates and keys now exist in "/etc/kubernetes/pki"
[kubeconfig] Wrote KubeConfig file to disk: "admin.conf"
[kubeconfig] Wrote KubeConfig file to disk: "kubelet.conf"
[kubeconfig] Wrote KubeConfig file to disk: "controller-manager.conf"
[kubeconfig] Wrote KubeConfig file to disk: "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 33.004078 seconds
[uploadconfig] Storing the configuration used in ConfigMap "kubeadm-config" in the "kube-system" Namespace
[markmaster] Will mark node localhost.master as master by adding a label and a taint
[markmaster] Master localhost.master tainted and labelled with key/value: node-role.kubernetes.io/master=""
[bootstraptoken] Using token: 2bc1d8.896c082ff3e4d6e6
[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: kube-dns
[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 --token 2bc1d8.896c082ff3e4d6e6 192.168.56.100:6443 --discovery-token-ca-cert-hash sha256:327b2050ded1cff763a20425d5df622abffb8d7148604b39e31c4c99b7e2110b

激动归激动，活得继续干~

注意上面输出的 join 命令行里面的 token 数据

kubeadm join --token 2bc1d8.896c082ff3e4d6e6 192.168.56.100:6443 --discovery-token-ca-cert-hash sha256:327b2050ded1cff763a20425d5df622abffb8d7148604b39e31c4c99b7e2110b

保存好，后面别的节点加入集群需要这个玩意~

初始化收尾

经过上面的步骤，k8s终于初始化完成了，但是现在 Master 节点的 kubectl 却还用不了：

[root@localhost ~]# kubectl get nodes
The connection to the server localhost:8080 was refused - did you specify the right host or port?

注意上面初始化完成的输出：

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/

是滴，我们还需要执行一点收尾工作

[root@localhost ~]# mkdir -p $HOME/.kube
[root@localhost ~]# cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
[root@localhost ~]# chown $(id -u):$(id -g) $HOME/.kube/config

再试试 kubectl 查看下节点：

[root@localhost ~]# kubectl get nodes
NAME               STATUS     ROLES     AGE       VERSION
localhost.master   NotReady   master    11m       v1.9.0

ok，初始化工作就算完成了！

设置 Master 为可调度节点

默认情况下，Master 不参与 Pod 调度，也就是说不会在 Master 节点上部署其他非系统 Pod。我们可以使用一下命令调整这个策略：

# 允许 Master 部署 Pod
kubectl taint nodes localhost.master node-role.kubernetes.io/master- --overwrite

# 禁止 Master 部署 Pod
kubectl taint nodes localhost.master node-role.kubernetes.io/master=:NoSchedule --overwrite

# 其中 localhost.master 是操作的节点名称

四、可能遇到的问题汇总

Docker和k8s的cgroup-driver配置不一致造成初始化失败

error: failed to run Kubelet: failed to create kubelet: misconfiguration: kubelet cgroup driver: "cgroupfs" is different from docker cgroup driver: "systemd"​

Docker 的 cgroup-driver 查看：

[root@local ~]# docker info | grep Cgroup
Cgroup Driver: cgroupfs

修改 kubelet 启动配置文件，主要是将 --cgroup-driver 改为cgroupfs（确保和 /usr/lib/systemd/system/docker.service 的用户一致就可以了，如果一致则不需要修改！）

[root@localhost ~]# vim /etc/systemd/system/kubelet.service.d/10-kubeadm.conf
[Service]
Environment="KUBELET_KUBECONFIG_ARGS=--kubeconfig=/etc/kubernetes/kubelet.conf --require-kubeconfig=true"
Environment="KUBELET_SYSTEM_PODS_ARGS=--pod-manifest-path=/etc/kubernetes/manifests --allow-privileged=true"
Environment="KUBELET_NETWORK_ARGS=--network-plugin=cni --cni-conf-dir=/etc/cni/net.d --cni-bin-dir=/opt/cni/bin"
Environment="KUBELET_DNS_ARGS=--cluster-dns=10.96.0.10 --cluster-domain=cluster.local"
Environment="KUBELET_AUTHZ_ARGS=--authorization-mode=Webhook --client-ca-file=/etc/kubernetes/pki/ca.crt"
Environment="KUBELET_CADVISOR_ARGS=--cadvisor-port=0"
Environment="KUBELET_CGROUP_ARGS=--cgroup-driver=cgroupfs"
ExecStart=
ExecStart=/usr/bin/kubelet $KUBELET_KUBECONFIG_ARGS $KUBELET_SYSTEM_PODS_ARGS $KUBELET_NETWORK_ARGS $KUBELET_DNS_ARGS $KUBELET_AUTHZ_ARGS $KUBELET_CADVISOR_ARGS $KUBELET_CGROUP_ARGS $KUBELET_EXTRA_ARGS

重启 docker 或 kubelet

systemctl daemon-reload
systemctl enable kubelet
systemctl restart kubelet

初始化 Preflight 检查警告

在执行初始化过程中,可能会输出一些警告和错误

[preflight] Running pre-flight checks.
	[WARNING Hostname]: hostname "localhost.master" could not be reached
	[WARNING Hostname]: hostname "localhost.master" lookup localhost.master on 8.8.8.8:53: no such host
	[WARNING Service-Kubelet]: kubelet service is not enabled, please run 'systemctl enable kubelet.service'
	[WARNING Firewalld]: firewalld is active, please ensure ports [6443 10250] are open or your cluster may not function correctly
	[WARNING FileExisting-crictl]: crictl not found in system path
[preflight] Some fatal errors occurred:
	[ERROR Swap]: running with swap on is not supported. Please disable swap
[preflight] If you know what you are doing, you can make a check non-fatal with `--ignore-preflight-errors=...`

如上面所示,出现 WARNING 不会终止初始化过程,出现 ERROR 就会终止初始化过程。例如上面就抛出没有关闭 Swap 页面分区的错误停止了初始化。
如果确实需要忽略所有 preflight 警告和错误，可以在 kubeadm init 命令后加上 --ignore-preflight-errors=... 的参数。

初始化查询版本文件超时

[root@localhost ~]# kubeadm init --pod-network-cidr="10.24.0.0/16"
unable to get URL "https://dl.k8s.io/release/stable-1.9.txt": Get https://storage.googleapis.com/kubernetes-release/release/stable-1.9.txt: dial tcp 172.217.160.112:443: i/o timeout

指定版本重新执行命令（这个文件目前内容是 v1.9.8）：

[root@localhost ~]# kubeadm init --kubernetes-version=1.9.8

[kubelet-check] It seems like the kubelet isn't running or healthy.

初始化过程下拉镜像步骤之后出现：

···
[init] This might take a minute or longer if the control plane images have to be pulled.
[kubelet-check] It seems like the kubelet isn't running or healthy.
[kubelet-check] The HTTP call equal to 'curl -sSL http://localhost:10255/healthz' failed with error: Get http://localhost:10255/healthz: dial tcp [::1]:10255: getsockopt: connection refused.
[kubelet-check] It seems like the kubelet isn't running or healthy.
[kubelet-check] The HTTP call equal to 'curl -sSL http://localhost:10255/healthz' failed with error: Get http://localhost:10255/healthz: dial tcp [::1]:10255: getsockopt: connection refused.
[kubelet-check] It seems like the kubelet isn't running or healthy.
[kubelet-check] The HTTP call equal to 'curl -sSL http://localhost:10255/healthz' failed with error: Get http://localhost:10255/healthz: dial tcp [::1]:10255: getsockopt: connection refused.
[kubelet-check] It seems like the kubelet isn't running or healthy.
[kubelet-check] The HTTP call equal to 'curl -sSL http://localhost:10255/healthz/syncloop' failed with error: Get http://localhost:10255/healthz/syncloop: dial tcp [::1]:10255: getsockopt: connection refused.
[kubelet-check] It seems like the kubelet isn't running or healthy.
[kubelet-check] The HTTP call equal to 'curl -sSL http://localhost:10255/healthz/syncloop' failed with error: Get http://localhost:10255/healthz/syncloop: dial tcp [::1]:10255: getsockopt: connection refused.
[kubelet-check] It seems like the kubelet isn't running or healthy.
[kubelet-check] The HTTP call equal to 'curl -sSL http://localhost:10255/healthz/syncloop' failed with error: Get http://localhost:10255/healthz/syncloop: dial tcp [::1]:10255: getsockopt: connection refused.
[kubelet-check] It seems like the kubelet isn't running or healthy.

原因是 Docker 和 k8s 的 cgroup-driver 不一致造成的（不知道为什么没有预检查出来）：
参考 5-1 【Docker和k8s的cgroup-driver配置不一致造成初始化失败】

iptables 拦截 bridge 网卡数据

初始化报错如下:

[init] Using Kubernetes version: v1.9.8
[init] Using Authorization modes: [Node RBAC]
[preflight] Running pre-flight checks.
	[WARNING Hostname]: hostname "localhost.master" could not be reached
	[WARNING Hostname]: hostname "localhost.master" lookup localhost.master on 8.8.8.8:53: no such host
	[WARNING FileExisting-crictl]: crictl not found in system path
[preflight] Some fatal errors occurred:
	[ERROR FileContent--proc-sys-net-bridge-bridge-nf-call-iptables]: /proc/sys/net/bridge/bridge-nf-call-iptables contents are not set to 1
[preflight] If you know what you are doing, you can make a check non-fatal with `--ignore-preflight-errors=...`

禁止 iptables 对 bridge 数据进行处理

cat <<EOF >  /etc/sysctl.d/k8s.conf
net.bridge.bridge-nf-call-ip6tables = 1
net.bridge.bridge-nf-call-iptables = 1
EOF
sysctl -p /etc/sysctl.d/k8s.conf

“按照”博客说的，还是初始化失败

尝试升级 Docker

emmm...最开始,我在部署 Docker 的时候,图方便就用 yum 安装的,版本是 1.13.2。然后在这期间试了很多次初始化(包括在国外服务器搭建，但环境一致),都是初始化失败...直到周一上班那天早上,无意间想准备一下 Docker 离线安装包，就去看了看 k8s v1.9.7 支持的 Docker 版本最高是 Docker-ce-17.03.x ,于是就卸载了原来的版本并安装了 17.03.2 版本的 Docker-ce —— 然后就初始化成功了...
虽然我也不确定是不是我用的 k8s 1.9.8 和 Docker 1.13.2 不兼容,但是大家要是遇到这种无解的情况,也可以尝试下升级 Docker。

最大的坑：防火墙

之前操作的时候，都没有关闭防火墙。直到我在写第四篇的时候，关掉了防火墙，才发现好多问题都是 Master 防火墙造成的。无语~

五、引用 & 鸣谢

  1. kubernetes(1.9.1部署)
  2. 安装部署 Kubernetes 集群
  3. 从零开始搭建Kubernetes集群
  4. kubeadm 搭建 kubernetes 集群
  5. kubeadm快速部署kubernetes1.6.7
  6. 阿里云快速部署Kubernetes - VPC环境

还有很多百度出来没摘抄的文章，感谢作者前辈们！