System requirements

System requirements for configuring GlusterFS.

Hardware requirements

You must use at least one dedicated node for GlusterFS.
The storage device that is used for GlusterFS must have a capacity of at least 25 GB.
The storage device that you use for GlusterFS must be a raw disk. It must not be formatted, partitioned, or used for file system storage needs.

The minimum CPU, Memory, RAM, and disk space requirements for the dedicated GlusterFS nodes are as shown in Table 1.

Table 1. Minimum hardware requirements for a GlusterFS node
Requirement	GlusterFS node	Notes
Number of hosts	1 or more
Cores	2 or more	Choose a modern processor with multiple cores. Servers often come packaged with two sockets (two processors), so the total number of cores in the system is the number of sockets multiplied by the number of cores in each processor. Today, processors typically have at least eight cores per processor. Higher end processors can have 20 or more cores per processor. When it comes to choosing a processor, it might be of best value to take a moderate approach. That is, select a processor that has an average (not the highest or the lowest) number of cores, frequency, and cache size.
CPU	>= 2.4 GHz	For a Linux® cluster, choose a CPU that supports SSE 4.2. For a Linux® on Power® (ppc64le) cluster, use a CPU that is version Power8 or higher.
RAM	>=8 GB	Each GlusterFS volume needs about 30 MB of RAM on each GlusterFS node. The total amount of RAM you need depends upon the number of volumes that might be required for your cluster.
Disk space to install	>=150 GB

IBM® Cloud Private Version 3.1.0 or later must be installed.
IBM Cloud Private supports IBM GlusterFS Chart v1.3.0, GlusterFS version 4.1.5, and Heketi version 8.0.0.
GlusterFS client must be installed on all the nodes that use GlusterFS volume.
dm_thin_pool kernel modules must be loaded on all storage nodes where GlusterFS servers are installed.

For more information about installing GlusterFS client and loading dm_thin_pool kernel modules, see Preparing the nodes.

Make sure that these ports are open but not in use by any service:

Table 1. GlusterFS ports
Port	Usage
24007	GlusterFS daemon
24008	GlusterFS management
2222	sshd (used when GlusterFS runs in a pod)
49152:49251	TCP port for each brick in a volume

Check if the ports are open by running the following command:

   netstat -an | grep <port number> | grep -i listen

If the command returns an output, it means that the port is open and is in use. Stop the service that is using the port.