38-40
Cisco Catalyst Blade Switch 3130 and 3032 for Dell Software Configuration Guide
OL-12247-04
Chapter 38 Configuring IP Unicast Routing
Configuring EIGRP
To create an EIGRP routing process, you must enable EIGRP and associate networks. EIGRP sends
updates to the interfaces in the specified networks. If you do not specify an interface network, it is not
advertised in any EIGRP update.
If you have routers on your network that are configured for IGRP, and you want to change to EIGRP, you
must designate transition routers configured with
both
IGRP and EIGRP. In these cases, perform Steps
1 through 3 in the next section and also see the
“Configuring Split Horizon” section on page 38-26
. You
must use the same autonomous-system number for routes so that they are automatically redistributed.
EIGRP Nonstop Forwarding
The switch stack supports two levels of EIGRP nonstop forwarding:
•
EIGRP NSF Awareness, page 38-40
•
EIGRP NSF Capability, page 38-40
EIGRP NSF Awareness
The IP-services feature set supports EIGRP NSF awareness for IPv4. When the neighboring router is
NSF-capable, the Layer 3 switch continues to forward packets from the neighboring router during the
interval between the primary route processor in a router failure and the backup RP take-over, or while
you manually reload the primary route processor for a nondisruptive software upgrade.
This feature cannot be disabled. For more information on this feature, see the “EIGRP Nonstop
Forwarding (NSF) Awareness” section of the
Cisco IOS IP Routing Protocols Configuration Guide,
Release 12.4
at this URL:
http://www.cisco.com/en/US/products/ps6350/products_configuration_guide_chapter09186a00804529
72.html
EIGRP NSF Capability
The IP-services feature set also supports EIGRP NSF-capable routing for IPv4 for better convergence
and lower traffic loss following a stack-master change. When an EIGRP NSF-capable stack master
restarts or a new stack master starts and NSF restarts, the switch has no neighbors, and the topology table
is empty. The switch must bring up the interfaces, re-acquire neighbors, and rebuild the topology and
routing tables without interrupting the traffic directed to the switch stack. EIGRP peer routers maintain
the routes learned from the new stack master and continue to forward traffic through the NSF restart
process.
To prevent an adjacency reset by the neighbors, the new stack master uses a new restart bit in the EIGRP
packet header. When the neighbor receives it, it synchronizes the stack in its peer list and maintains the
adjacency with the stack. The neighbor then sends its topology table to the stack master with the restart
bit set to show that it is NSF-aware and is aiding the new stack master.
If at least one of the stack peer neighbors is NSF-aware, the stack master receives updates and rebuilds
its database. Each NSF-aware neighbor sends an end-of-table (EOT) marker in the last update packet to
mark the end of the table content. The stack master recognizes the convergence when it receives the EOT
marker and begins sending updates. When the stack master has received all EOT markers from its
neighbors or when the NSF converge timer expires, EIGRP notifies the routing information database
(RIB) of convergence and floods its topology table to all NSF-aware peers.
Note
NSF is not supported on interfaces configured for Hot Standby Router Protocol (HSRP).
