glusterfs配置实例

目录

• 1 简介
• 2 配置4个服务端存储节点
• 2.1 第一个存储节点: server0.vol
• 2.2 第二个存储节点: server1.vol
• 2.3 第三个存储节点: server2.vol
• 2.4 第四个节点，命名空间节点: server3.vol
• 3 配置客户端节点: client.vol
• 4 启动存储集群
• 5 进一步阅读

简介

介绍配置glusterfs模拟存储集群:

• 一个挂载点
• 三个服务端存储节点，一个命名空间节点

配置4个服务端存储节点

为每个存储服务端建立一个目录:

$ sudo mkdir /home/cfs{0,1,2,3}

$ sudo chmod 1777 /home/cfs{0,1,2,3}

$ ls -ld /home/cfs[0-3]

drwxrwxrwt  2 root root 4096 2007-01-14 13:42 /home/cfs0

drwxrwxrwt  2 root root 4096 2007-01-14 13:42 /home/cfs1

drwxrwxrwt  2 root root 4096 2007-01-14 13:42 /home/cfs2

drwxrwxrwt  2 root root 4096 2007-01-14 13:42 /home/cfs3

然后为每个服务端编写一个vol配置文件，使之监听不同端口。

第一个存储节点: server0.vol

volume brick

        type storage/posix

        option directory /home/cfs0

end-volume

volume server

        type protocol/server

        option transport-type tcp/server

        option listen-port 6996

        option bind-address 127.0.0.1

        subvolumes brick

        option auth.ip.brick.allow 127.0.0.1

end-volume

第二个存储节点: server1.vol

volume brick

        type storage/posix

        option directory /home/cfs1

end-volume

volume server

        type protocol/server

        option transport-type tcp/server

        option listen-port 6997

        option bind-address 127.0.0.1

        subvolumes brick

        option auth.ip.brick.allow 127.0.0.1

end-volume

第三个存储节点: server2.vol

volume brick

        type storage/posix

        option directory /home/cfs2

end-volume

volume server

        type protocol/server

        option transport-type tcp/server

        option listen-port 6998

        option bind-address 127.0.0.1

        subvolumes brick

        option auth.ip.brick.allow 127.0.0.1

end-volume

第四个节点，命名空间节点: server3.vol

volume brick

        type storage/posix

        option directory /home/cfs3

end-volume

volume server

        type protocol/server

        option transport-type tcp/server

        option listen-port 6999

        option bind-address 127.0.0.1

        subvolumes brick

        option auth.ip.brick.allow 127.0.0.1

end-volume

配置客户端节点: client.vol

volume client0

 type protocol/client

 option transport-type tcp/client

 option remote-host 127.0.0.1

 option remote-port 6996

 option remote-subvolume brick

end-volume

volume client1

 type protocol/client

 option transport-type tcp/client

 option remote-host 127.0.0.1

 option remote-port 6997

 option remote-subvolume brick

end-volume

volume client2

 type protocol/client

 option transport-type tcp/client

 option remote-host 127.0.0.1

 option remote-port 6998

 option remote-subvolume brick

end-volume

volume ns

 type protocol/client

 option transport-type tcp/client

 option remote-host 127.0.0.1

 option remote-port 6999

 option remote-subvolume brick

end-volume

volume bricks

  type cluster/unify

  subvolumes client0 client1 client2

  namespace ns

  option rr.limits.min-free-disk 10GB # In later versions, its changed to % based value.

  option scheduler rr

end-volume

### Add writebehind feature

volume writebehind

  type performance/write-behind

  option aggregate-size 131072 # unit in bytes

  subvolumes bricks

end-volume

### Add readahead feature

volume readahead

  type performance/read-ahead

  option page-size 65536     # unit in bytes

  option page-count 16       # cache per file  = (page-count x page-size)

  subvolumes writebehind

end-volume

启动存储集群

启动4个服务监听各自端口，守护各自目录:

$ sudo glusterfsd -f ./server0.vol

$ sudo glusterfsd -f ./server1.vol

$ sudo glusterfsd -f ./server2.vol

$ sudo glusterfsd -f ./server3.vol

$ ps w -C glusterfsd

  PID TTY      STAT   TIME COMMAND

13524 ?        Ss     0:00 [glusterfsd]

13526 ?        Ss     0:00 [glusterfsd]

13528 ?        Ss     0:00 [glusterfsd]

13532 ?        Ss     0:00 [glusterfsd]

然后通过客户端挂载:

$ mkdir ./glusterfs

$ sudo glusterfs -f ./client.vol ./glusterfs

检查是否挂载成功:

$ mount | tail -1

glusterfs:13548 on /home/toad/gfs/glusterfs type fuse (rw,allow_other,default_permissions)

说明挂载成功！

进一步阅读

• GlusterFS_Utility_Storage_Cluster Storage cluster is an utility I wrote to quickly set up clusters of X servers and 1 local client node. Its in Perl, and its ugly, it was not supposed to be a release of any kind.