1.2. Prerequisites

Ensure that your environment meets the following requirements:

Important

For production use and high availability, a minimum of three servers per trusted storage pool is required.
For quorum reasons, multi-site clusters with synchronous replication require an odd number of sites (minimum of three) to survive a site failure without manual intervention.

1.2.1. Multi-site Cluster Latency

Ensure the latency of the multi-site cluster is no more than 5ms. Execute the following command on each site to determine the latency with the other site:
# ping -c100 -q site_ip_address/site_hostname

Note

For more information, see KCS#413623.

1.2.2. File System Requirements

XFS - Format the back-end file system using XFS for glusterFS bricks. XFS can journal metadata, resulting in faster crash recovery. The XFS file system can also be defragmented and expanded while mounted and active.

Note

Red Hat assists existing Gluster Storage Software Appliance customers using ext3 or ext4 to upgrade to a supported version of Red Hat Gluster Storage using the XFS back-end file system.

1.2.3. Logical Volume Manager

Format glusterFS bricks using XFS on the Logical Volume Manager to prepare for the installation.

1.2.4. Network Time Configuration

Synchronize time across all Red Hat Gluster Storage servers using the Network Time Protocol (NTP) or Chrony daemon. For more information about these features see Choosing Between NTP Daemons

1.2.4.1. Configuring time synchronization using Chrony

Note

For Red Hat Enterprise Linux 8, chrony is enabled by default and Network Time Protocol (NTP) is no longer supported.
For information on configuring chrony for time synchronization for all Gluster nodes, see Migrating to chrony.
After configuring chrony, verify that time is synchronized for all Gluster nodes by using the following command from the primary node to all other nodes:
# clockdiff node-hostname

1.2.4.2. Configuring time synchronization using Network Time Protocol

For Red Hat Enterprise Linux 7
Use a remote server over the Network Time Protocol (NTP) to synchronize the system clock. For more information on NTP, refer to Configuring NTP Using ntpd. Set the ntpd daemon to automatically synchronize the time during the boot process as follows:
  1. Edit the NTP configuration file /etc/ntp.conf using a text editor such as vim or nano.
    # nano /etc/ntp.conf
  2. Add or edit the list of public NTP servers in the ntp.conf file as follows:
    server 0.rhel.pool.ntp.org
    server 1.rhel.pool.ntp.org
    server 2.rhel.pool.ntp.org
    The Red Hat Enterprise Linux 6 version of this file already contains the required information. Edit the contents of this file if customization is required. For more information regarding supported Red Hat Enterprise Linux version for a particular Red Hat Gluster Storage release, see Section 1.7, “Red Hat Gluster Storage Support Matrix”.
  3. Optionally, increase the initial synchronization speed by appending the iburst directive to each line:
    server 0.rhel.pool.ntp.org iburst
    server 1.rhel.pool.ntp.org iburst
    server 2.rhel.pool.ntp.org iburst
  4. After the list of servers is complete, set the required permissions in the same file. Ensure that only localhost has unrestricted access:
    restrict default kod nomodify notrap nopeer noquery
    restrict -6 default kod nomodify notrap nopeer noquery
    restrict 127.0.0.1
    restrict -6 ::1
  5. Save all changes, exit the editor, and restart the NTP daemon:
    # service ntpd restart
  6. Ensure that the ntpd daemon starts at boot time:
    # chkconfig ntpd on
Use the ntpdate command for a one-time synchronization of NTP. For more information about this feature, see the Red Hat Enterprise Linux Deployment Guide.