Appendix

HA cluster in place upgrade

HA Cluster in place upgrades are the preferred choice when you must work exclusively with the machines you already have provisioned. It is important to note that this approach is recommended only when provisioning new machines or nodes is not feasible (contrast this to the preferred Autopilot upgrade recommended above).

It is worth emphasizing that HA upgrades demand strict adherence to a well-defined sequence of steps(opens in new tab). Deviating from the prescribed order can lead to complications, especially if standby nodes are not replaced before the active node. Such missteps can disrupt the upgrade process and potentially introduce vulnerabilities.

DR upgrade

Autopilot upgrades are only for use with integrated storage. Migrating from another storage backend to integrated storage would best be done with a DR-style upgrade.

Disaster recovery upgrades are necessary in situations that require a failover mechanism. This process involves replicating all data to new nodes, temporarily halting incoming traffic, promoting the DR secondary to the primary role, and subsequently reconfiguring the nodes to communicate with the DR secondary. The primary objective is to ensure seamless system continuity and minimal downtime during unexpected failures or disruptions.