Standard And Extended Esrp - Extreme Networks ExtremeWare XOS Guide Manual

Displaying ESRP-Aware Information
To display ESRP-aware information, use the following command:
show esrp {<name>}
The display includes the group number and MAC address for the master of the group, as well as the
age of the information.

Standard and Extended ESRP

ESRP has two modes of operation: standard and extended. By default, ExtremeWare XOS operates in
extended mode. To configure a different mode of operation, use the following command:
configure esrp mode [extended | standard]
Standard mode is backward compatible with and supports the ESRP functionality of switches running
ExtremeWare. ESRP functionality available in extended mode is not applicable in standard mode. Use
standard mode if your network contains both switches running ExtremeWare and switches running
ExtremeWare XOS participating in ESRP.
Extended mode supports and is compatible with switches running ExtremeWare XOS while
participating in ESRP. Use extended mode if your network contains only switches running ExtremeWare
XOS.
The following list describes the major differences in behavior between standard and extended mode:
Handshaking
●
In standard mode, events such as link flapping cause the ESRP master switch to generate a large
number of packets and to increase processing time.
To prevent this, extended mode supports handshaking. Handshaking occurs when a switch requests
a state change, forces its neighbor to acknowledge the change, and the neighbor sends an
acknowledgement to the requesting switch. For example, if a slave switch wants to become the
master, it enters the pre-master state, notifies the neighbor switch, and forces the neighbor to
acknowledge the change. The neighbor then sends an acknowledgement back to the slave switch.
While the requesting switch waits for the acknowledgements, future updates are suppressed to make
sure the neighbor does not act on incorrect data.
Stickiness
●
In standard mode, if an event causes the ESRP master switch to fail over to the slave, it becomes the
new master. If another event occurs, the new master switch returns to the slave and you have
experienced two network interruptions.
To prevent this, extended mode supports the sticky election metric. The default election algorithm
uses the sticky metric. For example, if an event causes the ESRP master switch to fail over to the
slave, it becomes the new master and has a higher sticky value. If another event occurs, for example
adding active ports to the slave, the new master does not fail back to the original master even if the
slave has more active ports. After sticky is set on the master, regardless of changes to its neighbor's
election algorithm, the new master retains its position. Sticky algorithms provide for fewer network
interruptions than non-sticky algorithms. Sticky is set only on the master switch.
Port weight
●
In standard mode, the port count calculation does not take into account the available bandwidth of
the ports. For example, a switch with a one Gigabit Ethernet uplink may be unable to become master
because another switch has a load-shared group of four fast Ethernet links. The active port count
only consider the number of active ports, not the bandwidth of those ports.
ExtremeWare XOS 11.1 Concepts Guide
ESRP Concepts
329