Product guide

Router Redundancy Using XRRP
Overview of XRRP Operation
If Communication is Maintained Through Non-XRRP Interfaces. In
some cases, it may be possible that all connectivity is lost between the routers
on all their XRRP virtual router interfaces, in which case XRRP operates and
both routers try to take control of all the virtual routers in the Protection
Domain, but if connectivity still exists on non-XRRP VLANs, a situation could
occur in which both routers allow and use the same MAC addresses on the
non-XRRP VLAN(s). This could create a situation in which a switch connected
between the two routers will see continuous move interrupts and potential
duplication of inbound packets if that switch floods. To prevent this condition,
a simple XRRP protocol packet is exchanged between the two routers on the
non-XRRP VLAN to inform each other of their uses of the MAC addresses. This
exchange prevents the routers from taking over each other’s MAC addresses.
Note that this protocol is used only when one router attempts to take over
control of the other router’s virtual router interfaces.
XRRP Infinite Fail-Back for the 5300xl Switches
Introduction
XRRP infinite failback is an optional enhancement to the Series 5300xl XRRP
routing feature, and is designed to reduce network disruption due to peer
router fail-backs occurring automatically. This is accomplished by configuring
XRRP with the infinite fail-back option and then using a manual fail-back
command that can be executed at the discretion of a system operator.
Overview of Infinite Fail-Back Operation
If a fail-over event occurs in an XRRP protection domain, the peer for the failed
router automatically takes over the routing function for the failed router. This
peer router, which is already the master for its own XRRP VLAN(s) and
XRRP MAC address (primary address control), also becomes the master for
the XRRP VLAN(s) and XRRP MAC address for the failed router (secondary
address control).
The Problem. Prior to software release E.09.05, if a failed XRRP router
recovers access to all of its XRRP VLANs, then a fail-back automatically
occurs, which removes secondary address control from the fail-back router
and restores control of these addresses to the recovered router. (The fail-back
router ceases to advertise XRRP packets for the failed router’s backed-up IP
addresses.) This automatic fail-back can cause a network slowdown due to a
disruption of the TCP connections during the fail-back. In cases where a fail-
over/fail-back cycle occurs repeatedly, frequent network disruptions can
occur.
12-11