What I usually do is set the hello time to 1sec, and the max
message age time and the forward delay time to 4 sec or so.

If you divide the max message age time by the hello time, you
get the maximum number of consecutive STP packets that can be
lost before a recalculation.  Be sure to adapt this to observed
packet loss.

Twice the forward delay time is the time it takes between a
recalculation and a new stable situation (in case of single
host failure).

So, max message age time plus twice the forward delay time
is the failover time.


On Mon, Jun 10, 2002 at 11:32:55AM -0700, Jeremy Junginger wrote:

> The bridging firewall with kernel 2.4.18 is running beautifully.  I've
> been running it on a test segment with a single host behind it in order
> to test out the rules and connectivity.  The next step would be to
> implement failover.  I have a pretty good understanding of the spanning
> tree protocol, bridge priority, root bridge, and the like, but would
> like to get your input on minimizing failover time without causing
> unnecessary STP recalculations.  Has anyone experimented with the STP
> timers to determine an optimal setting for quick failover?  Just
> curious...
> 
> -Jeremy


_______________________________________________
Bridge mailing list
[EMAIL PROTECTED]
http://www.math.leidenuniv.nl/mailman/listinfo/bridge

Reply via email to