Written on June 3, 2012 at 11:17 am by

Install & Configure GlusterFS -CentOS / RHEL

Filed under Servers {no comments}




GlusterFS is a software-only, highly available, scalable, centrally managed storage pool.

Installing GlusterFS

Server Installation
Download & Install rpm from

OR build rpm from tar
Install Dependency

yum -y groupinstall "Development tools"
yum -y install fuse fuse-devel rpcbind readline-devel libibverbs-devel rpm-devel

download latest tar from http://download.gluster.com/pub/gluster/glusterfs/LATEST/
Build RPM

rpmbuild -ta glusterfs-3.2.5.tar.gz
cd ./rpmbuild/RPMS/x86_64
rpm -ivh * / yum -y install *.rpm

Enable service on boot & start

chkconfig glusterd on
service glusterd start

Client Installation

yum -y install glusterfs-core* glusterfs-fuse* glusterfs-rdma*

Server Configuration

Here we are using following two server ip & client
LAN 192.168.2.0/23
192.168.2.92 client1
192.168.3.132 gserver1
192.168.3.141 gserver2

Probe peer server, run below command on server1 to probe server2

#on gserver1
gluster peer probe 192.168.3.141
#get peer / server status
gluster peer status
#create volume
mkdir /storage
gluster volume create storevol replica 2 transport tcp 192.168.3.132:/storage 192.168.3.141:/storage
Creation of volume storevol has been successful. Please start the volume to access data.
 
#storevol is volume name used
#replica 2 is for two bricks to be configured as replica, you can add more bricks later but atleast two at a time.
#here /storage mount point is used to store GlusterFS files / folders
#start volume
gluster volume start storevol
Starting volume storevol has been successful
#get volume info
gluster volume info
#By default, all clients can connect to the volume. If you want to grant access to client2 ( 192.168.2.102)
gluster volume set testvol auth.allow 192.168.2.102

Client Configuration

mkdir /home1
mount -t glusterfs 192.168.3.132:/storevol /home1/

Here we have mounted glusterfs from server1 to mount point home1.
Now even if server1 fails, client can access files stored on /home1 via server2. There is no need to remount from another server etc. While mounting from server1, client fetches volume info for other servers too. Hence no mount point / server ip change required.

Files / Folders copied on /home1 will be replicated on both servers.
Incase of server1 failed, files will be stored on server2 which is synced back to server1 on restore.

Add / Remove bricks from GlusterFS

While adding / removing bricks you need to add / remove atleast two bricks. if single is used following error will appear.

gluster volume add-brick storevol 192.168.2.104:/storage

Incorrect number of bricks supplied 1 for type REPLICATE with count 2

Additional servers for demo
192.168.2.104 server3
192.168.2.223 server4

#add peer 
gluster peer probe 192.168.2.104
gluster peer probe 192.168.2.223
#add brick to gluster
gluster volume add-brick storevol 192.168.2.104:/storage 192.168.2.223:/storage

BELOW COMMAND WILL CAUSE DATA LOSS.

#remove brick
gluster volume remove-brick storevol 192.168.2.223:/storage 192.168.2.104:/storage

References
Manage Bricks
Building GlusterFS RPM
Howtoforge


if you find any missing point in here, please let us know in comment section or tweet us at @linuxreaders. To get more articles like this, subscribe to our RSS feeds / Mails.
Read 240 articles by

Trending Posts