前几天给大家介绍了docker的集群存储软件etcd的安装（地址http://dl528888.blog.51cto.com/2382721/1623746），今天就再给大家介绍集群的控制软件fleet安装。

fleet的介绍，info里复制的

fleet 是一个通过 Systemd对CoreOS 集群中进行控制和管理的工具。fleet 与 Systemd 之间通过 D-Bus API 进行交互，每个 fleet agent 之间通过 etcd 服务来注册和同步数据。fleet 提供的功能非常丰富，包括查看集群中服务器的状态、启动或终止 Docker container、读取日志内容等。

简单来说，我搭建docker集群，使用etcd作为信息存储，fleet连接与控制所有节点服务器的systemd，然后通过想用的命令创建或者消灭节点里docker容器。

下面是我自己画的一个fleet的架构图

fleet与etcd都是coreos开发，内置到coreos系统里，但我后来在弄docker集群的时候，在centos7系统里部署与使用这2个软件，使用起来很方便。

下面是安装

1、安装go

yum install go -y

2、下载fleet

cd /tmp/

git clone https://github.com/coreos/fleet.git

3、安装fleet

cd fleet

./build

4、负责可执行程序

cp bin/fleetd /usr/local/bin/

cp bin/fleetctl /usr/local/bin/

5、配置fleet

在某一个节点里编辑与配置/etc/fleet/fleet.conf

09:15:02 # cat /etc/fleet/fleet.conf

# This config file is INI-formatted



# Lower the logging threshold. Acceptable values are 0, 1, and 2. A higher

# value corresponds to a lower logging threshold.

# verbosity=0



# Provide a custom set of etcd endpoints. The default value is determined

# by the underlying go-etcd library.

#etcd_servers=["http://192.168.65.2:4001"]



# Amount of time in seconds to allow a single etcd request before considering it failed.

# etcd_request_timeout=1.0



# Provide TLS configuration when SSL certificate authentication is enabled in etcd endpoints

# etcd_cafile=/path/to/CAfile

# etcd_keyfile=/path/to/keyfile

# etcd_certfile=/path/to/certfile



# IP address that should be published with any socket information. By default,

# no IP address is published.

public_ip="10.10.27.221"



# Comma-delimited key/value pairs that are published to the fleet registry.

# This data can be referenced in unit files to affect scheduling descisions.

# An example could look like: metadata="region=us-west,az=us-west-1"

metadata="region=cn-core-02"



# An Agent will be considered dead if it exceeds this amount of time to

# communicate with the Registry. The agent will attempt a heartbeat at half

# of this value.

# agent_ttl="30s"



# Interval at which the engine should reconcile the cluster schedule in etcd.

# engine_reconcile_interval=2

基本只需要修改public_ip与metadata

6、配置启动脚本

09:16:14 # cat /usr/lib/systemd/system/fleet.service;

[Unit]

Description=Fleet

Documentation=https://github.com/coreos/fleet



[Service]

Type=simple

EnvironmentFile=-/etc/fleet/fleet.conf

ExecStart=/usr/local/bin/fleetd -config /etc/fleet/fleet.conf

ExecStop=/usr/bin/killall -9 fleetd

Restart=on-failure

RestartSec=5s

[Install]

WantedBy=multi-user.target

7、启动fleet

systemctl start fleet

8、查看集群状态

我目前已经创建好了集群，并且有了4个节点

09:15:06 # fleetctl list-machines -l

MACHINE					IP		METADATA

1a1b60030aaef2f1d8d0468f0000002d	10.10.21.199	region=cn-core-01

6d832c54a40fe752dba35d0b00000024	10.10.27.221	region=cn-core-02

893b160e363b4ec7834719a7f06e67cf	10.10.17.3	region=cn-core-04

cce83e1dd55b46f6a543b49a6b59586a	10.10.17.3	region=cn-core-03

其中machine是一个主机唯一的标示，ip是之前配置fleet.conf里的public_ip项，metadata是fleet.conf里metadata的。

目前已经建立好集群，请直接安装fleet前一定要先在每个节点安装etcd，否则fleet无法工作的，之后给大家整体介绍一个docker集群的真实案例。