For a stack to operate it must have a master unit. A master unit is the active unit that handles the configuration of the stack, while the other units assume the slave role. Additionally, a unit within the stack also assumes the role of master backup, in the event that the master unit fails.
The objective of this document is to understand unit failures in a stack and the resulting master backup process on the SG550XG and SG350XG Series Managed Switches.
For a full length demonstration of Master Backup, please view the video below:
Suppose the stack is in a ring topology, with Unit 1 as the master unit, Unit 2 as the backup master, and Units 3 and 4 as slave units. If the master is disconnected at all points from the ring or chain topology, Unit 2 will assume the new master role.
This process of the backup unit taking over the master role is called a switchover. When the switchover occurs, the backup unit becomes the master, and all of its processes and protocols are initialized to take responsibility for the entire stack. As a result, there is temporarily no traffic forwarded to this unit; however, the slave units remain active.
When the backup unit becomes the master, the active slave units remain active and continue to forward packets based on the configuration from the original master, minimizing data traffic interruption. Once the backup unit has completed the transition to the master state, it initializes the slave units one at a time by resetting the configuration of the slave unit to default. This helps prevent any incorrect configurations from the new master unit.
The master backup process is an efficient way for a stack to continue to function normally in the event of a unit failure. If the master unit fails, the backup unit assumes the master role and continues to function as the master. This switchover prevents the failure of the master from causing the entire stack to fail. The original master unit can also be reconnected and become the master unit once again, while the backup master unit can resume its original backup role as well. This allows for easy topology changes within the stack, without disrupting the rest of the slave units.