2 master election, 3 stack redundancy – PLANET SGS-5220-24P2X User Manual
Page 340
User’s Manual of SGS-5220 Series
General Switch ID Assignment Rules
When assigning Switch IDs to the devices in the stack, you must note the following:
1.
Switches with assigned IDs can be changed to use any other switch ID (possibly by swapping Switch ID with
another active switch).
2.
When swapping two Switch IDs, the devices will retain their (own) configuration, except for the Switch ID.
3.
Switches without an assigned Switch ID can only be assigned to any unused ID.
4.
When assigning a Switch ID of an inactive switch to a new switch, the new switch will inherit the former's
configuration (see "Replacing a Switch" above).
5.
Deleting a switch will remove any configuration pertaining to it.
6.
Deleting an active switch will leave it with an unassigned Switch ID until rebooted or manually assigning a Switch
ID.
4.19.1.2 Master Election
Within a managed stack, one master switch (or just "master") must be elected. Any switch not being master is a slave switch (or
just "slave").
To elect a master, the following criteria are evaluated sequentially:
1. If any switch already claims to have been master for more than 30 seconds, then that switch will become
master.
2. If multiple switches claim to have been master for more than 30 seconds, then the switch which has been
master for the longest period of time will become master.
3. The switch with the smallest master priority.
4. The switch with the smallest MAC address.
The above algorithm ensures that once a master has been elected and has been master for more than 30 seconds, it will remain
master. However in some cases the user may want to enforce a new master election.
4.19.1.3 Stack Redundancy
In the unlikely event that a SGS-5220 Series Switch fails in a stack, stack integrity is maintained if the redundant cable is
connected to the stack. The affected switch within the sack can be replaced or removed without disrupting normal operation.
The broken link is bypassed and data transmission continues uninterrupted. The single management IP address for the stack is
also preserved for uninterrupted management and monitoring.
340