Dell Brocade DCX, Руководство по настройке

Страница: 472 / 1436

400 BigIron RX Series Configuration Guide
53-1001810-01
Ring initialization for shared interfaces
14
For example, in Figure 71 , the ID of all interfaces on all nodes on Ring 1 is 1 and all interfaces on
all nodes on Ring 2 is 2. Port 1/1 on node S1 and Port 2/2 on S2 have the IDs of 1 and 2 since the
interfaces are shared by Rings 1 and 2.
The ring’s ID is also used to determine an interface’s priority. Generally, a ring’s ID is also the ring’s
priority and the priority of all interfaces on that ring. However, if the interface is shared by two or
more rings, then the highest priority (lowest ID) becomes the priority of the interface. For example,
in
Figure 71 , all interfaces on Ring 1, except for Port 1/1 on node S1 and Port 2/2 on node S2 have
a priority of 1. Likewise, all interfaces on Ring 2, except for Port 1/1 on node S1 and Port 2/2 on
node S2 have a priority of 2. Port 1/1 on S1 and Port 2/2 on S2 have a priority of 1 since 1 is the
highest priority (lowest ID) of the rings that share the interface.
If a node has interfaces that have different IDs, the interfaces that belong to the ring with the
highest priority become regular ports. Those interfaces that do not belong to the ring with the
highest priority become tunnel ports. In
Figure 71 , nodes S1 and S2 have interfaces that belong to
Rings 1 and 2. Those interfaces with a priority of 1 are regular ports. The interfaces with a priority
of 2 are the tunnel ports since they belong to Ring 2, which has a lower priority than Ring 1.
How ring breaks Are detected and healed between shared interfaces
If the link between shared interfaces breaks, the secondary interface on Ring 1’s master node
changes to a preforwarding state. The RHP packet sent by port 3/1 on Ring 2 is forwarded through
the interfaces on S4, then to S2. The packet is then forwarded through S2 to S3, but not from S2 to
S1 since the link between the two nodes is not available. When the packet reaches Ring 1’s master
node, the packet is forwarded through the secondary interface since it is currently in a
preforwarding state. A secondary interface in preforwarding mode ignores any RHP packet that is
not from its ring. The secondary interface changes to blocking mode only when the RHP packet
forwarded by its primary interface is returned.
The packet then continues around Ring 1, through the interfaces on S1 to Ring 2 until it reaches
Ring 2’s master node. Port 3/2, the secondary interface on Ring 2 changes to blocking mode since
it received its own packet, then blocks the packet to prevent a loop.
NOTE
On the ring member node, the primary and secondary interface is not only decided by the
configuration, but also decide by the RHP flow from the ring master. The primary and secondary
interface may not be swapped even if the configuration changes and there is an active ring master
in the topology. If there is no active ring master in the topology, then the running configuration of the
interface on the member node will follow what was configured.
Selection of master node
Allowing MRP rings to share interfaces limits the nodes that can be designated as the master node.
Any node on an MRP ring that does not have a shared interface can be designated as the ring’s
master node. However, if all nodes on the ring have shared interfaces, nodes that do not have
tunnel ports can be designated as the master node of that ring. If none of the nodes meet these
criteria, you must change the rings’ priorities by reconfiguring the rings’ ID.
In Figure 71 , any of the nodes on Ring 1, even S1 or S2, can be a master node since none of its
interfaces are tunnel ports. However in Ring 2, neither S1 nor S2 can be a master node since these
nodes contain tunnel ports.