How to create a GlusterFS volume

Creating a GlusterFS volume turns multiple server directories into a single storage namespace that can be mounted by clients and expanded over time. Volume creation is where capacity and redundancy choices become permanent, so a clean layout now avoids fragile fixes later.

A GlusterFS volume is built from bricks, which are export directories hosted by peers in a trusted storage pool. The glusterd management daemon stores the volume definition and propagates it across the pool, while clients fetch a volume configuration file and connect directly to the participating brick servers.

Stable name resolution and firewall access are required between all nodes before volume creation, because management traffic uses fixed ports while brick I/O consumes ports from a configurable range. Brick directories should be empty and placed on dedicated storage (commonly XFS) instead of the operating system filesystem, and replica layouts should include quorum planning to reduce split-brain risk during network partitions.

1. Create a trusted storage pool for GlusterFS.

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

2. Confirm that each peer hostname resolves to the expected address.

   $ getent hosts node1 node2
   10.0.0.11   node1
   10.0.0.12   node2

   Unresolvable peer names commonly break gluster probe, volume create, and client mounts.

3. Verify that all peers show a connected state in the trusted storage pool.

   $ sudo gluster peer status
   Number of Peers: 1
   
   Hostname: node2
   Uuid: 8b1b8c1e-0e1b-4a0f-9c7d-0b6d9bb0d5c2
   State: Peer in Cluster (Connected)

4. Allow network access for required GlusterFS ports on all nodes.

   $ sudo firewall-cmd --zone=public --add-port=24007-24008/tcp --permanent
   success
   $ sudo firewall-cmd --zone=public --add-port=24007-24008/udp --permanent
   success
   $ sudo firewall-cmd --zone=public --add-port=49152-60999/tcp --permanent
   success
   $ sudo firewall-cmd --reload
   success

   Replace public with the active firewalld zone when needed.
   Brick ports are allocated from base-port to max-port (commonly 49152 to 60999) with one port consumed per brick.
   Older releases used 24009+ instead of 49152+, while Gluster 10+ can randomize brick ports within the configured range in /etc/glusterfs/glusterd.vol.

5. Create directory for the GlusterFS brick on each node.

   $ sudo mkdir -p /var/data/gluster/brick

   For production, place each brick on a dedicated partition or block device instead of the operating system filesystem.

   XFS is the recommended filesystem type for GlusterFS volumes.

   Related: How to mount disks and partitions in Linux

6. Create the GlusterFS volume from the first node.

   $ sudo gluster volume create volume1 replica 2 transport tcp node1:/var/data/gluster/brick node2:/var/data/gluster/brick force
   volume create: volume1: success: please start the volume to access data

   Replicated volumes require each replica brick on a different peer, even when brick paths match across servers.
   The force option bypasses brick placement checks and is best avoided unless the brick path is confirmed empty on dedicated storage.

   Two-way replication (replica 2) can enter split-brain during network partitions; prefer an arbiter or three-way replica layout when consistency must be preserved.

   Related: Types of GlusterFS volumes

7. Start the newly created volume from the first node.

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

8. Verify the GlusterFS volume details from the first node.

   $ sudo gluster volume info all
   
   Volume Name: volume1
   Type: Replicate
   Volume ID: 19550419-3495-45d7-bdc6-cab4fa4fb516
   Status: Started
   Snapshot Count: 0
   Number of Bricks: 1 x 2 = 2
   Transport-type: tcp
   Bricks:
   Brick1: node1:/var/data/gluster/brick
   Brick2: node2:/var/data/gluster/brick
   Options Reconfigured:
   cluster.granular-entry-heal: on
   storage.fips-mode-rchecksum: on
   transport.address-family: inet
   nfs.disable: on
   performance.client-io-threads: off

9. Verify that all bricks are online from the first node.

   $ sudo gluster volume status volume1
   
   Status of volume: volume1
   Gluster process                             TCP Port  RDMA Port  Online  Pid
   ------------------------------------------------------------------------------
   Brick node1:/var/data/gluster/brick         49152     0          Y       2147
   Brick node2:/var/data/gluster/brick         49152     0          Y       2099
   Self-heal Daemon on node1                   N/A       N/A        Y       2318
   Self-heal Daemon on node2                   N/A       N/A        Y       2284
   
   Task Status of Volume volume1
   ------------------------------------------------------------------------------
   There are no active volume tasks