On PowerStore, each storage resource is assigned to either node A or node B for load-balancing and redundancy purposes. Besides storage-resource assignments, each node has various containers running on them that make up the PowerStore operating system. If one node becomes unavailable, its resources (storage and containers) automatically failover to the surviving node.

The time that it takes for the failover process to complete depends on several factors such as system utilization and the number of resources. The peer node assumes ownership of the resources and continues servicing I/O to avoid an extended outage. Failovers occur if the following occurs on a node:

• Node reboot: The system or a user rebooted the node.
• Hardware or software fault: The node has failed and must be replaced.
• Service mode: The system or a user placed the node into service mode. This occurs automatically when the node is unable to boot due to a hardware or software issue.
• Powered off: A user powered off the node.

Note: Manually putting a node into service mode is only available through a service script. It is not available from the PowerStore Manager, REST API, or PSTCLI.

While the node is unavailable, all the resources of the node are serviced by the peer. After the node is brought back online or the fault is corrected, block-storage resources automatically fail back to the proper node owner. File-storage resources must be failed back manually.

During a code upgrade, both nodes reboot in a coordinated manner. All resources on the rebooting node are failed over to the peer node. When the peer comes back online, the resources are failed back to their original owner. This process repeats for the second node. Users can run a pre-upgrade health check before starting a code upgrade to ensure a smooth upgrade process.