Storage array subsystems need code upgrades as fixes, patches, or feature upgrades. You can perform these upgrades online when the filesystem is mounted and I/Os are being served to the storage.
Legacy storage subsystems contain two controllers for redundancy. This upgrade is done one controller at a time, and DMP will failover all I/O to the second controllers when the first controller is undergoing an Online Controller Upgrade. After the first controller has completely staged the code, it reboots/resets and comes up online with the new version of the code. The second controller goes through the same process and I/O will failover to the first controller.
Array vendors have different names for this process. For example, EMC calls it a nondisruptive upgrade (NDU) for Clariion arrays.
There is no special handling required for A/A type arrays during this process. For A/P and A/PF and ALUA type arrays, there is array-specific handling which DMP needs to do via vendor-specific array policy modules (APMs) during an online controller code upgrade.
When a controller resets itself, goes down, and reboots during a code upgrade, DMP senses this state via SCSI Status and will immediately failover all I/O to the next controller.
To verify which arrays support Online Controller Upgrade or NDU, see the hardware compatibility list (HCL) at the following URL: