Detailed explanation of kubernetes (IV) -- kubernetes deployment based on kubedm

Today I will continue to introduce Linux Operation and maintenance related knowledge , The main content of this paper is based on kubeadm Of Kubernetes Deploy .

One 、Kubeadm brief introduction

kubeadm yes Kubernetes The cluster construction tool provided by the cluster , utilize Kubeadm Can quickly build a Kubernetes colony , And can achieve Kubernetes Cluster upgrade 、 Degradation and removal . utilize Kubeadm Way to deploy Kubernetes In the cluster , Most of the resources are Pod How to run .
Kubeadm Official address: ：https://kubernetes.io/docs/reference/setup-tools/kubeadm, Its page is as follows ：

Kubeadm The project is hosted to Github On , Its address is ：https://github.com/kubernetes/kubeadm, The page is as follows ：

Two 、 Preparation before installation

Next , We will use it Kubeadm To deploy Kubernetes colony .
our Kubernetes The cluster installation and deployment environment is as follows ：
192.168.136.15 Master
192.168.136.14 Node1
192.168.136.12 Node2
We modified the of three devices /etc/hosts file , Make its host name and IP The address corresponds to .
Next , We carry out Kubernetes Preparations before cluster installation and deployment .
On the above three nodes , Let's install Docker.
First , Let's first install some dependent environments and auxiliary tools , Carry out orders ：

yum -y install yum-utils device-mapper-persistent-data lvm2

In the software installed above ,yum-utils Provides yum Management commands for yum-config-manager,device-mapper-persistent-data yes Linux2.6 General device mapping mechanism supporting logical volume management in kernel , It provides a highly modular kernel architecture for the implementation of block device driver for storage resource management .
after , We configure our YUM Warehouse , Make it point to Alibaba cloud , Carry out orders ：

yum-config-manager --add-repo https://mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo

Next , We can install it Docker 了 , Carry out orders ：

yum -y install docker-ce-18.09.7 docker-ce-cli-18.09.7 containerd.io

Can finish Docker Installation .
install Docker After completion , Because we are running in a virtual machine scenario , Hardware resources are in short supply , So we need to modify Cgroup The driver , Change it to systemd, This can make the server node more stable under the premise of resource shortage .
open /etc/docker/daemon.json file （ If not, create ）, Enter the following in this file ：

{
    
        "exec-opts":["native.cgroupdriver=systemd"]
}

After completion , We have to restart Docker, Carry out orders ：

systemctl restart docker

Then let's look at Cgroup drive , Carry out orders ：

docker info | grep Cgroup

The results are shown below ：

You can see , We have successfully modified Docker The driver .

3、 ... and 、kubeadm install

Next , Let's install Kubeadm.
First , Let's configure Kubernetes Of YUM Warehouse , ad locum , We use alicloud's YUM Warehouse . Create a new one /etc/yum.repo.d/kubernetes.repo file , And write the following contents in the file ：

[kubernetes]
name=Kubernetes
baseurl=https://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64/
enabled=1
gpgcheck=1
repo_gpgcheck=0

After completion , You can install kubeadm 了 , Carry out orders ：

yum -y install kubelet-1.15.2 kubeadm-1.15.2 kubectl-1.15.2 --nogpgcheck

such , our kubeadm The installation is complete , The reason YUM The command is followed by –nogpgcheck, It's because we don't configure verification .
Because we use virtual machines for configuration , In order to make Kubernetes The cluster can run normally , Therefore, it is possible to use SWAP Partition , At this time , We are Master When initializing on , You cannot configure disable SWAP, So you have to configure to ignore SWAP Partition error reporting .
open /etc/sysconfig/kubelet, Modify the contents of the file as follows ：

KUBELET_EXTRA_ARGS="--fail-swap-on=false"

that will do .

Four 、Master Node initialization

Next , We need to be in Master Initialization on node Kubernetes colony . stay Master On the node , Carry out orders ：

kubeadm init --kubernetes-version=v1.15.2 --image-repository registry.aliyuncs.com/google_containers --pod-network-cidr=10.10.0.0/16 --service-cidr=10.20.0.0/12 --ignore-preflight-errors=Swap

In the above order ,–kubernetes-version The parameter specifies Kubernetes Version of ;–image-repository The parameter specifies Kubernetes The image download address of the cluster , Here, due to the network , We Kubernetes The official image of cannot be used , We can only use the image of domestic Alibaba cloud ;–pod-network-cidr The parameter specifies pod Of the resource object IP Address segment ;–service-cidr Parameter specifies service Of the resource object IP Address segment ;–ignore-preflight-errors Parameter specifies ignore Swap Partition error reporting .
The specified execution takes a long time , The results are shown below ：

You can see , The command was executed successfully , At the end of the execution result, we give some operations we need to do , as well as Node Commands for nodes to join the cluster .
We do the above , Carry out orders ：

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

After execution , Let's check the currently downloaded image , Carry out orders ：

docker images

The results are as follows ：

after , We have to download flannel Components , Download the component first , Download at ：https://raw.githubusercontent.com/coreos/flannel/master/Documentation/kube-flannel.yml
Carry out orders ：

wget https://raw.githubusercontent.com/coreos/flannel/master/Documentation/kube-flannel.yml

Download the file to your local , And then execute the command ：

kubectl apply -f kube-falnnel.yml

The results are as follows ：

such ,Master The node configuration is completed .

5、 ... and 、Node Node join

Next , We have to control Node Node join Kubernetes colony . We copy directly in Master Display after cluster initialization , The display is to join Kubernetes Cluster command . however , Because we enabled Swap The problem of zoning , So we need to add –ignore-preflight-errors=Swap, therefore , Carry out orders ：

kubeadm join 192.168.136.15:6443 --token zc7nhx.96pky3iq2qwiztox     --discovery-token-ca-cert-hash sha256:0cadb638b11d86854ea85114cc1c900c28490a49cc8d17c1cd9583dac1fa1dc6 --ignore-preflight-errors=Swap

The execution result of this command is as follows ：

You can see , our Node Join the cluster successfully ！

6、 ... and 、 Effect display

Last , Let's test the effect of our configuration .
stay Master And two Node After all nodes are configured , We are Master On the node , Carry out orders ：

kubectl get nodes

give the result as follows ：

You can see , After a period of time , our Master and Node1、Node2 The node is working normally .
Originality is not easy. , Reprint please explain the source ：https://blog.csdn.net/weixin_40228200