Table of ContentsView in Frames

How controller failover works

Data ONTAP is equipped with functionality called controller failover that allows the partner system to assume responsibility for the failed system's LUNs without interruption.

A storage system with an HA pair has a single global WWNN, and both systems in the HA pair function as a single FC node. Each node in the HA pair shares the partner node's LUN map information.

All LUNs in the HA pair are available on all ports in the HA pair by default. As a result, there are more paths to LUNs stored on the HA pair because any port on each node can provide access to both local and partner LUNs. You can specify the LUNs available on a subset of ports by defining port sets and binding them to an igroup. Any host in the igroup can access the LUNs only by connecting to the target ports in the port set.

The following figure shows an example configuration with a multi-attached host. If the host accesses lun_1 through ports 4a, 4b, 5a, or 5b on storage system X, then storage system X recognizes that lun_1 is a local LUN. If the host accesses lun_1 through any of the ports on storage system Y, lun_1 is recognized as a partner LUN and storage system Y sends the SCSI requests to storage system X over the HA pair interconnect.