How to create a dispersed GlusterFS volume

Dispersed GlusterFS volumes provide fault tolerance with lower raw-capacity overhead than full replication, making them a strong fit when usable space matters as much as availability.

A dispersed volume uses erasure coding to split each file into data and parity fragments across a disperse set of bricks. The redundancy value determines how many bricks in that set can be lost while the volume stays online, and Disperse Data represents the number of data fragments (total bricks minus redundancy).

Redundancy must be greater than 0 and the total number of bricks must be greater than 2 * redundancy (minimum three bricks), and a disperse set should avoid multiple bricks on the same peer to prevent single-node failures from removing multiple fragments at once. All bricks in a disperse set should have the same capacity because the smallest brick caps usable space, and omitting an explicit redundancy value can trigger an interactive prompt during volume creation.

1. Confirm all peers are connected in the trusted storage pool.

   $ sudo gluster peer status
   Number of Peers: 5
   
   Hostname: node2
   Uuid: 0b6c2f2d-7b7e-4f3f-9d4a-2f3e5b7d1a9c
   State: Peer in Cluster (Connected)
   
   Hostname: node3
   Uuid: 6a4f5c3d-1a2b-4c5d-9e0f-1a2b3c4d5e6f
   State: Peer in Cluster (Connected)
   ##### snipped #####

   Related: How to set up a trusted storage pool for GlusterFS

2. Create the brick directory on each node that will host a brick.

   $ sudo mkdir -p /srv/gluster/brick1

   Brick directories must be empty and should not be nested inside another brick path.

   Use a dedicated filesystem mounted under /srv/gluster for production bricks to avoid filling the system volume.

3. Create the dispersed volume with the chosen brick count using an explicit redundancy level.

   $ sudo gluster volume create volume1 disperse 6 redundancy 2 transport tcp node1:/srv/gluster/brick1 node2:/srv/gluster/brick1 node3:/srv/gluster/brick1 node4:/srv/gluster/brick1 node5:/srv/gluster/brick1 node6:/srv/gluster/brick1
   volume create: volume1: success: please start the volume to access data

   A 6-brick disperse set with redundancy 2 stores 4 data fragments plus 2 parity fragments per stripe.

   Using force to place multiple bricks from the same disperse set on one peer reduces fault tolerance for node failures.

4. Start the new volume.

   $ sudo gluster volume start volume1
   volume start: volume1: success

   Volumes must be started before client mounts to avoid hung client operations.

5. Verify the volume reports the expected disperse count with the intended redundancy.

   $ sudo gluster volume info volume1
   Volume Name: volume1
   Type: Disperse
   Status: Started
   Number of Bricks: 1 x 6 = 6
   Disperse Data: 4
   Redundancy: 2
   ##### snipped #####

6. Verify every brick is online for the started volume.

   $ sudo gluster volume status volume1
   Status of volume: volume1
   Gluster process                             TCP Port  RDMA Port  Online  Pid
   ------------------------------------------------------------------------------
   Brick node1:/srv/gluster/brick1             49152     0          Y       1712
   Brick node2:/srv/gluster/brick1             49153     0          Y       1730
   Brick node3:/srv/gluster/brick1             49154     0          Y       1748
   Brick node4:/srv/gluster/brick1             49155     0          Y       1764
   Brick node5:/srv/gluster/brick1             49156     0          Y       1781
   Brick node6:/srv/gluster/brick1             49157     0          Y       1799
   ##### snipped #####