GlusterFS
GlusterFS is a cluster file-system capable of scaling to several peta-bytes. It aggregates various storage bricks over Infiniband RDMA or TCP/IP interconnect into one large parallel network file system. GlusterFS is based on a stackable user space design without compromising performance. For a list of features, see GlusterFS Features.
Contents |
[edit] Installation
[edit] FUSE support (prerequisites)
Filesystems in Userspace (aka FUSE) is required in order to be able to mount GlusterFS in a local directory. If this functionality is desired, be sure to have fuse enabled in your kernel and sys-fs/fuse installed for the required userland tools, as well as the "fuse" USE-flag set on sys-cluster/glusterfs.
[edit] Installing Gluster
Just unmask (if necessesary) and emerge glusterfs:
[edit] Running a simple Gluster setup
The first thing to do is to get the init script for glusterd working. You will need a simple configuration file for glusterd, like the following:
volume management type mgmt/glusterd option working-directory /var/lib/glusterd option transport-type socket,rdma option transport.socket.keepalive-time 10 option transport.socket.keepalive-interval 2 end-volume
With glusterd running, you can now use the "gluster" command to create, activate and configure volumes.
First, runIf one of these does not list your glusterfs, then something went wrong, and you should check the log file specified in the mount command that you just used. Also try restarting glusterd.
If they do list the mounted glusterfs, then everything should be all right, and you should now be able use your local one-machine "cluster" file-system through the directory it is mounted on. Now (optionally) grab some coffee, and then refer to the #External links in order to configure an actual production volume, glusterfsd, add more machines to the cluster, and so on.