AP hearbeat timers and fast failover

[Nigel Bowden]

I read about the introduction of the fast heartbeat parameter (in 5.0) that can be configured from a WLC CLI to provider faster fail-over times for APs when their primary WLC fails.

I can't find a doc that clearly explains how the setting works, or its relationship with the existing heartbeat timer.

For instance, if I set the timers like this :

config advanced timers ap-heartBeat-timeout 3

config advanced timers ap-fast-heartbeat all enable 1

...what is the effect?

The ap-heartBeat-timeout has to be 3 times the value of the ap-fast-heartbeat timer, but why?

Does this mean that the AP's will failover in 3 seconds (i.e. a hot/warm fail-over), or will they detect the WLC loss in 3 seconds, then recover after their standard reboot?

Any light shed on this subject would be gratefully received.

Thanks in advance

Nigel.

[dennischolmes · Accepted solution]

I am not a fan of faster heartbeats as this can cause your APs to reboot when you don't want them too. What this does is in effect increase the heartbeat timers that require the AP to hear packets broadcasting the availability of a controller. Normally heartbeats are run at a much slower time so as to not negatively effect the network with excessive broadcasts. When the APs senses it has missed heartbeats from a controller it will accelerate the heartbeat to make sure that no collision or errant network problem prohibited the heartbeat from being heard. If the heartbeats are still not heard then the AP reboots and begins the discovery process and searched for master, primary, secondary, tertiary, and least congested available controllers. So, in essence all you are doing is speeding up the rediscovery process.