Jim,

Thanks a million for the quick reply! If I discover how to modify the
polling time period, I'll be sure to post it.

--
Anthony


"James Egan" wrote:

>
> On Tue, 24 Jun 2008 12:18:01 -0700, Anthony
> <Anthony@discussions.microsoft.com> wrote:
>
> >route -p add 192.168.1.0 mask 255.255.255.0 192.168.1.253 metric 1
> >route -p add 192.168.1.0 mask 255.255.0.0 192.168.1.254 metric 10
>
> I tested something similar myself recently except my test involved two
> default gateways out to the Intenet. The primary with a manual metric
> of 1 and the secondary with a manual metric of 20.
>
> Although I left the primary router powered up, I zapped the dsl
> synchronisation effectively killing the gateway.
>
> >
> >The data is very time-sensitive, so I cannot have much failover latency. If
> >my primary route (through .253) goes down,
> >
> >A) How long will it take to fail over to the backup?
>
> Only a few seconds.
>
> >
> >B) Will XP then ignore the primary route and consider the backup as the
> >primary, or will XP keep checking to see if the primary is up every so often
> >(how often)?
>
> I can't say exactly how often but in my test the default gateway
> automatically switched back to using the one with the lower metric
> every few minutes even though it was still dead and swapped again back
> to the secondary gateway on detecting it was still dead.
>
> So it appears that it follows the same procedure each time and polls
> the primary gateway for a response every few minutes.
>
> >
> >C) If my primary goes back up again, will traffic then shift back to it?
>
> My suspicion is that it will happen after the next poll whenever it is
> due. I don't know if this polling period is settable in the registry
> somewhere.
>
> Overall I would say that it was an annoyance for me getting the
> occasional waits for switching between gateways. It certainly wasn't
> transparent. I could tell something wasn't quite right but it could be
> a lot more serious for you and you may have to look for a better
> solution.
>
>
> Jim.
>
>