Configuring OSPF
own forwarding engine that uses this information to make forwarding decisions locally on the
module that receives the frame. These engines independently make forwarding decisions based
on route and rule information distributed by the router protocol process. In a stable network, the
distributed route and rule information is fairly constant. If the router protocol process was to
suddenly fail, forwarding information current at the time of the failure in all probability is usable
for the short time after the failure until recovery occurs. During this recovery period, existing
connections (that were not directly using the failed module) remain in effect. New connections
continue to be installed using the last known ʺgoodʺ forwarding information. The router protocol
process that failed is dynamically restarted on another module. The user does not configure where
the router process is running. The router forwarding process remains active on every module. The
protocol process exchanges protocol and maintains state that it distributes to the other modules
and does not have to run on any specific module. One exception to this rule is that the module
must have 256M of memory to be router protocol process eligible.
Upon failure of a module running the router protocol process, the protocol process is started on a
recovery module. One of the first messages it sends to its OSPF neighbors is a grace LSA. High
availability failover will successfully occur if the following is true:
•
The router is enabled for graceful restart
•
The neighbors are enabled to participate as graceful restart helper
•
The OSPF dead interval is configured for a sufficient period such that the grace LSA is
received by its neighbors before the configured OSPF dead interval expires
•
And each neighbor is a member of a LAG common to the failed router, allowing the neighbor
to remain up
Figure 21-1
Figure
21‐1 depicts the physical and logical configurations of the single router high availability
failover mechanism. The blue lines display direct neighbor connections to the router enabled for
OSPF graceful restart and members of LAGs common to the failing router. The red lines display
VLAN connections common to both the failing and recovery routers.
Purpose
To enable and configure the Open Shortest Path First (OSPF) routing protocol.
21-20 Routing Protocol Configuration
Physical and Logical Single Router HA Failover Configuration
Neighbor A
1
2
3
4
5
6
7
1
2
1
2
3
4
5
6
7
Neighbor B
100.1.1.0/24
100.1.1.3
Router
3
4
5
6
7
100.1.1.0/24
VLAN 100
100.1.1.5
Neighbor A
Neighbor B
OSPF
LAG
Router
VLAN 100
redistribute