[c-nsp] SLA route tracking

[Arie Vayner (avayner)]

I would also recommend using the "delay up NN" option - this will allow
you to make sure the route is back and stays up, and not react to
network flapping.

Arie

-----Original Message-----
From: cisco-nsp-bounces at puck.nether.net
[mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of Peter Rathlev
Sent: Tuesday, January 18, 2011 20:36
To: Sergey Voropaev
Cc: cisco-nsp at puck.nether.net
Subject: Re: [c-nsp] SLA route tracking

On Tue, 2011-01-18 at 20:26 +0300, Sergey Voropaev wrote:
> We have following config on the 3750 switch.
> 
> track 10 rtr 10 reachibility
> !
> ip route 1.1.1.2 255.255.255.255 vlan10 1.1.1.2 track 10
> !
> ip sla 10
>  tcp-connect 1.1.1.2 33333 source-ip 1.1.1.1 source-port 433
> !
>  
> The main disadvantage of such config is that route removed from
> routing table in case of ONE TCP timeout. In this case false alarm
> occurs. Is it possible using 3750 platfirm remove route fo THREE
> CONSEQUINTIVE TCP timout?

You could use the "delay down" command on the track object, like this:

ip sla 10
 tcp-connect 10.20.30.40 22 control disable
 timeout 5000
 frequency 10
!
ip sla schedule 10 life forever start-time now
!
track 10 ip sla 10
 delay down 25
!

The SLA object probes every 10 seconds, but the track object only goes
down 25 seconds later (2,5 timers SLA frequency here). So if the SLA
object misses one or two probes and then regains connectivity, the track
will never go down.

I tried this on 12.2(55)SE IP Base; since you didn't specify your IOS
version I can't say if it's supported there. This was also on a C3560G,
but I can't imagine they're different in that aspect.

-- 
Peter


_______________________________________________
cisco-nsp mailing list  cisco-nsp at puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/