Allied Telesis SwitchBlade x3100 Series Manual page 451

SuperLoop Prevention
4.6.10.1 Overview of EPSR SuperLoop Ring Port Priority Configuration
To prevent a SuperLoop condition from occurring, the concept of certain ring interfaces having a specified pri-
ority is used.
This priority is assigned to the interfaces that make up the common link shared by EPSR domains. The value
range is 0 to 127. By default, the priority of each of the ring interfaces for an EPSR domain is 0 (the lowest pri-
ority). The higher values, however, are used when there are interconnected EPSR rings in which the SuperLoop
condition needs to be avoided.
To use the priorities, the user should take the following steps:
Review the network topologies and see which ring segments have conditions in which the SuperLoop condi-
1.
tion could occur.
Assign these ring segments a unique priority number. The user should start with 127 for the ring segment
2.
that is closest to the upstream network and go down in intervals of 1 (127, 126, 125..) for the ring segments
as they go farther away form the upstream network.
Assign the interfaces for the common link with the priority of the ring segment.
3.
Assign the interfaces for the master node with the priority of the ring segment.
4.
The following figures show the result of taking these steps.
Figure 4-13 shows a sample interconnected topology where priorities have been configured. Also, a nested ring
has been included. Ring 1 is closest to the upstream network and is therefore assigned priority 127. For the
control VLAN (V_60), this priority has been assigned to the master node interfaces and the common link inter-
faces. Ring 2 has priority 126 and its control VLAN (V_90) has this priority assigned to its master node and
common link interfaces. Ring 3 (the nested loop within ring segment 2) has priority 125 and its control VLAN
(V_40) has this priority assigned to its master node and common link interfaces.
Figure 4-14 shows what happens when there is a fault with the common link. Since ring segment 1 has the
higher priority on the interface, it will unblock its SP interface, but the lower priority ring segments 2 and 3 will
not. As a result, the SuperLoop condition is prevented.
Caution: As long as a fault exists in the common link, the lower priority masters will continue to block their SP
interfaces, even if another fault occurs in the high priority ring. This could result in loss of service for
some nodes in the lower priority rings.
Software Reference for SwitchBlade x3100 Series Switches (Layer Two Switching)
Introduction
4-131