ETCD集群部署文档

ETCD部署及解释

一、部署环境

主机账号	主机密码
root	XXXXXXXXX

二、资源配置

|主机名|IP地址|CPU(核心)|内存(GB)|系统硬盘(GB)|挂载硬盘(GB)|
|-|-|-|-|-|-|-|
|etcd-k8s-master-1|192.168.40.140|2|4|20|100|
|etcd-k8s-master-2|192.168.40.141|2|4|20|100|
|etcd-k8s-master-3|192.168.40.142|2|4|20|100|

三、安装、启动etcd

3.1 安装ansible

# 安装ansible
yum install ansible -y

# 配置ansible hosts文件
vim /etc/ansible/hosts
[etcd]
192.168.40.140
192.168.40.141
192.168.40.142
[all:vars]
ansible_ssh_user="root"
ansible_ssh_pass="XXXXXXXXX"

# 编写分发文件
vim ssh-key.yml 
- hosts: all
  become: yes
  tasks: 
   - name: Auth
     authorized_key: 
      user: root
      key: "{{lookup('file','~/.ssh/id_rsa.pub')}}"

#生成ssh-key文件
ssh-keygen

3.2 编辑分发hosts文件

# 编辑hosts文件
vim /etc/hosts
127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4
::1         localhost localhost.localdomain localhost6 localhost6.localdomain6

192.168.40.140 etcd-1 etcd-k8s-master-1
192.168.40.141 etcd-2 etcd-k8s-master-2
192.168.40.142 etcd-3 etcd-k8s-master-3

# 分发hosts文件
ansible all -m copy -a "src=/etc/hosts dest=/etc/hosts"

3.2 安装golang

# 下载golang并创建对应目录
cd /opt
wget https://dl.google.com/go/go1.13.4.linux-amd64.tar.gz
tar xvf go1.13.4.linux-amd64.tar.gz go1.13.4.linux-amd64.tar.gz
mkdir -P /opt/gopath/{src,bin,pkg}

# 配置golang环境变量
vim /etc/profile.d/go.sh
    export GOROOT=/opt/go
    export GOPATH=/opt/gopath
    export PATH=$PATH:$GOROOT/bin

# 刷新环境变量
. /etc/profile.d/go.sh

3.2 编译etcd与etcdctl代码生成二进制文件与分发

3.2.1 编译生成与分发etcd

go get github.com/etcd-io/etcd
cd /opt/gopath/src/go.etcd.io/etcd
go build
ansible all -m copy -a "src=./etcd dest=/bin/etcd mode=u+rwx,g+rx,o+rx"

3.2.2 编译生成与分发etcd

cd /opt/gopath/src/go.etcd.io/etcd/etcdctl
go build
ansible all -m copy -a "src=./etcdctl dest=/bin/etcdctl mode=u+rwx,g+rx,o+rx"

此处如果发生报错：build go.etcd.io/etcd/etcdctl: cannot load github.com/bgentry/speakeasy: github.com/bgentry/[email protected]: Get https://proxy.golang.org/github.com/bgentry/speakeasy/@v/v0.1.0.zip: dial tcp 216.58.200.49:443: i/o timeout

执行:

git clone https://github.com/bgentry/speakeasy.git /opt/gopath/src/github.com/

3.3 etcd启动项

etcd节点1

nohup etcd --name etcd-1 \
  --initial-advertise-peer-urls 'http://192.168.40.140:2380' \
  --listen-peer-urls 'http://192.168.40.140:2380' \
  --advertise-client-urls 'http://192.168.40.140:2379' \
  --listen-client-urls 'http://192.168.40.140:2379' \
  --initial-cluster-token 'etcd-1' \
  --initial-cluster etcd-1=http://192.168.40.140:2380,etcd-2=http://192.168.40.141:2380,etcd-3=http://192.168.40.142:2380 \
  --data-dir /opt/${HOSTNAME}.etcd \
  --initial-cluster-state 'new'&

etcd节点2

nohup etcd --name etcd-2 \
  --initial-advertise-peer-urls 'http://192.168.40.141:2380' \
  --listen-peer-urls 'http://192.168.40.141:2380' \
  --advertise-client-urls 'http://192.168.40.141:2379' \
  --listen-client-urls 'http://192.168.40.141:2379' \
  --initial-cluster-token 'etcd-1' \
  --initial-cluster etcd-1=http://192.168.40.140:2380,etcd-2=http://192.168.40.141:2380,etcd-3=http://192.168.40.142:2380 \
  --data-dir /opt/${HOSTNAME}.etcd \
  --initial-cluster-state 'new'&

etcd节点3

nohup etcd --name etcd-3 \
  --initial-advertise-peer-urls 'http://192.168.40.142:2380' \
  --listen-peer-urls 'http://192.168.40.142:2380' \
  --advertise-client-urls 'http://192.168.40.142:2379' \
  --listen-client-urls 'http://192.168.40.142:2379' \
  --initial-cluster-token 'etcd-1' \
  --initial-cluster etcd-1=http://192.168.40.140:2380,etcd-2=http://192.168.40.141:2380,etcd-3=http://192.168.40.142:2380 \
  --data-dir /opt/${HOSTNAME}.etcd \
  --initial-cluster-state 'new'&

3.4 etcd启动项信息

• --name：etcd在集群中的节点名称，在同一个集群中必须是唯一的
• --initial-advertise-peer-urls：建议用于集群内部几点之间的交互的URL地址，几点见讲义该值进行通信
• --listen-peer-urls：用于集群内各个节点之间通信的URL地址，每个节点可以监听多个url地址,集群内部将通过这些URL地址进行数据交互，例如，leader几点的选举、Message消息的传输或者快照的传输等。
• --advertise-client-urls：建议客户端使用的RUL地址，该值用于etcd代理或etcd成员与etcd几点通信
• --listen-client-urls：用于当前地址与客户端交互的URL地址，每个几点可以向客户端提供多个URL地址
• --initial-cluster-token：集群的唯一标识
• --initial-cluster：集群中所有的initial-advertise-peer-urls
• --data-dic：数据的存放位置，${name}该节点的名字
• --initial-cluster-state new：新集群的标识

测试集群否正常

[root@etcd-k8s-master-1 ~]# etcdctl --endpoints=192.168.40.140:2379,192.168.40.141:2379,192.168.40.142:2379 endpoint health
192.168.40.140:2379 is healthy: successfully committed proposal: took = 2.268118ms
192.168.40.141:2379 is healthy: successfully committed proposal: took = 2.396264ms
192.168.40.142:2379 is healthy: successfully committed proposal: took = 3.042977ms

[root@etcd-k8s-master-1 ~]# etcdctl --write-out=table --endpoints=192.168.40.140:2379,192.168.40.141:2379,192.168.40.142:2379 endpoint status
+---------------------+------------------+-----------+---------+-----------+------------+-----------+------------+--------------------+--------+
|     ENDPOINT        |        ID        |  VERSION  | DB SIZE | IS LEADER | IS LEARNER | RAFT TERM | RAFT INDEX | RAFT APPLIED INDEX | ERRORS |
+---------------------+------------------+-----------+---------+-----------+------------+-----------+------------+--------------------+--------+
| 192.168.40.140:2379 | 1f5cb9a09297a825 | 3.5.0-pre |   20 kB |     false |      false |         9 |          9 |                  9 |        |
| 192.168.40.141:2379 | 8a5428646f2d78bf | 3.5.0-pre |   20 kB |      true |      false |         9 |          9 |                  9 |        |
| 192.168.40.142:2379 | f16ab1eb94b31b0d | 3.5.0-pre |   16 kB |     false |      false |         9 |          9 |                  9 |        |
+---------------------+------------------+-----------+---------+-----------+------------+-----------+------------+--------------------+--------+