High Availability (HA) Best Practices

  1. Use a dedicated witness for each HA cluster.

  2. With HA witness being a VM, be sure it is not running on the datastore using the same BrickStor HA shared storage.

  3. Use two or more resource groups to get more performance out of the BrickStor, making it active/active.

  4. Use an LACP 802.1ad aggregate for the data network across two or more stacked network switches. This will boost network performance by load balancing traffic across multiple ports and improve availability.

  5. Avoid manually failing over Resource Groups with pools in the degraded state. Degraded pools can take longer to import during failover and can result in a self induced outage. Resolve pool issues first and only then fail over the Resource Group.

  6. Avoid using jumbo frames. Jumbo frames can boost performance for data transportation. However, in NAS solutions it only fits in very specific environments and must be properly configured on all network devices. Improper use of jumbo frames can result in poor performance.

  7. Avoid using DNS hostnames for HA configuration. This eliminates dependency on DNS services.