Chinaunix首页 | 论坛 | 博客
  • 博客访问: 128860
  • 博文数量: 17
  • 博客积分: 10
  • 博客等级: 民兵
  • 技术积分: 207
  • 用 户 组: 普通用户
  • 注册时间: 2011-12-07 09:39
文章分类

全部博文(17)

文章存档

2017年(1)

2015年(1)

2014年(15)

我的朋友

分类: 服务器与存储

2014-08-28 09:58:08

 

安装

存储端

 

yum安装

wget -P /etc/yum.repos.d 

# yum -y install fuse  fuse-libs glusterfs glusterfs-rdma glusterfs-server glusterfs-devel glusterfs-fuse glusterfs-extra-xlators

 

验证fuse安装:

# modprobe fuse                         //载入fuse模块

# dmesg | grep -i fuse                  //看到版本号就OK

 

源码安装

#tar zxvf tar zxvfglusterfs-3.2.5.tar.gz

# cdglusterfs-3.2.5

#./configure  --prefix=/  //加上prefix配置文件在/etc下,不加会在/usr/local/etc

#make&& make install

#ldconfig                  //如果不执行会报错:libglusterfs.so.0 找不到。          
#glusterfs -V              //
查看版本,看是否安装完成。

 

 

启动停止服务

/etc/init.d/glusterd start

/etc/init.d/glusterd stop

 

服务自启动

 

Red Hat-based Systems

chkconfig glusterd on

 

Debian-based Systems

update-rc.d glusterd defaults

 

Systems Other than Red Hat and Debain

echo "/etc/init.d/glusterd start" >> /etc/rc.local

创建存储池

添加节点到存储池

# glusterpeer probe SERVERNAME/IP

 

For example

# gluster peer probe g2

Probe successful

# gluster peer probe g3

Probe successful

# gluster peer probe g4

Probe successful

 

察看状态

# gluster peer status

 

Number of Peers: 3

 

Hostname: g2

Uuid: 0fd210c8-60d5-4ef2-9120-e5f23923fbfe

State: Peer in Cluster (Connected)

 

Hostname: g3

Uuid: cf76a23b-11d4-4553-925e-6f7244091332

State: Peer in Cluster (Connected)

 

Hostname: g4

Uuid: 02418e07-0bd9-4f86-8d4c-e485837648e5

State: Peer in Cluster (Connected)

 

删除节点

# glusterpeer detach server

 

For example:

# gluster peer detach g4

Detach successful

创建卷

卷分为以下几种类型:

 

Distributed

Replicated

Striped

Distributed Striped

Distributed Replicated

Distributed Striped Replicated

Striped Replicated

命令格式:

# gluster volume create NEW-VOLNAME [stripe COUNT] [replica COUNT]

[transport [tcp | rdma | tcp,rdma]] NEW-BRICK...   

For example:创建 DistributedStriped Replicated Volumes 

# gluster volume create DSR-volume stripe 2 replica 2 transport tcp g1:/DSR1

g2:/DSR2 g3:/DSR3 g4:/DSR4 g1:/DSR5 g2:/DSR6 g3:/DSR7 g4:/DSR8  

# 文件将会被分割为2份 stripe ,其中存储在g1:/DSR1和 g2:/DSR2 中,另外一个存储在 g3:/DSR3和g4:/DSR4 中,其中g1:/DSR1 和 g2:/DSR2 中的内容各自互为备份;g3:/DSR3 和g4:/DSR4 中的内容各自互为备份 

# gluster volume info 

VolumeName: DSR-volume

Type:Distributed-Striped-Replicate

VolumeID: d27a7519-4a3c-444b-933f-d166924b94be

Status:Created

Numberof Bricks: 2 x 2 x 2 = 8 #第一个2代表distribute的数量;第二个2代表

stripe的数量;第三个则代表replica的数量。

Transport-type:tcp

Bricks:

Brick1:g1:/DSR1

Brick2:g2:/DSR2

Brick3:g3:/DSR3

Brick4:g4:/DSR4

Brick5:g1:/DSR5

Brick6:g2:/DSR6

Brick7:g3:/DSR7

Brick8:g4:/DSR8 

启动/停止卷

# gluster volume start VOLNAME

 

Forexample:

# glustervolume start/stop DSR-volume 

# 卷在启动后client才可以访问使用。 



阅读(1744) | 评论(0) | 转发(0) |
0

上一篇:GlusterFs 简介

下一篇:访问GlusterFs

给主人留下些什么吧!~~