IP SLA Tracking Objects Dropping

[ap_cisco0]

Hi,

I use tracking objects aroung the organization where I work to monitor WAN and VPN connections and add/remove routes based on the state of the object.  I'm having 2 locations that are constanty going up and down and I've been troubleshooting and monitoring for the last few weeks without finding anything.  I've been incrementing the timeout for the SLA and it seems like this is helping a little (less overall drops) but the drops still occur.  Our ISP reports no issues and we see no issues internally on the circuits.  Just out of curiosity could this be some kind of IOS bug or hardware malfunction?  The router logs are full of these:

Mar 21 16:18:33: %TRACKING-5-STATE: 2 ip sla 2 reachability Up->Down

Mar 21 16:18:38: %TRACKING-5-STATE: 2 ip sla 2 reachability Down->Up

Mar 21 17:24:14: %TRACKING-5-STATE: 2 ip sla 2 reachability Up->Down

Mar 21 17:24:19: %TRACKING-5-STATE: 2 ip sla 2 reachability Down->Up

Mar 21 17:32:14: %TRACKING-5-STATE: 2 ip sla 2 reachability Up->Down

Mar 21 17:32:19: %TRACKING-5-STATE: 2 ip sla 2 reachability Down->Up

Mar 22 05:02:42: %TRACKING-5-STATE: 2 ip sla 2 reachability Up->Down

Mar 22 05:02:47: %TRACKING-5-STATE: 2 ip sla 2 reachability Down->Up

Mar 22 05:42:23: %TRACKING-5-STATE: 1 ip sla 1 reachability Up->Down

Mar 22 05:42:28: %TRACKING-5-STATE: 1 ip sla 1 reachability Down->Up

Mar 22 12:49:25: %TRACKING-5-STATE: 2 ip sla 2 reachability Up->Down

Mar 22 12:49:30: %TRACKING-5-STATE: 2 ip sla 2 reachability Down->Up

Mar 22 14:45:36: %TRACKING-5-STATE: 2 ip sla 2 reachability Up->Down

Mar 22 14:45:41: %TRACKING-5-STATE: 2 ip sla 2 reachability Down->Up

Mar 22 15:18:46: %TRACKING-5-STATE: 2 ip sla 2 reachability Up->Down

Mar 22 15:18:51: %TRACKING-5-STATE: 2 ip sla 2 reachability Down->Up

Mar 22 15:22:36: %TRACKING-5-STATE: 2 ip sla 2 reachability Up->Down

Mar 22 15:22:41: %TRACKING-5-STATE: 2 ip sla 2 reachability Down->Up

Mar 22 15:28:46: %TRACKING-5-STATE: 2 ip sla 2 reachability Up->Down

Mar 22 15:28:51: %TRACKING-5-STATE: 2 ip sla 2 reachability Down->Up

Mar 22 15:57:26: %TRACKING-5-STATE: 2 ip sla 2 reachability Up->Down

Mar 22 15:57:31: %TRACKING-5-STATE: 2 ip sla 2 reachability Down->Up

Mar 24 21:50:48: %TRACKING-5-STATE: 2 ip sla 2 reachability Up->Down

Mar 24 21:50:53: %TRACKING-5-STATE: 2 ip sla 2 reachability Down->Up

Mar 25 11:14:17: %TRACKING-5-STATE: 2 ip sla 2 reachability Up->Down

Mar 25 11:14:22: %TRACKING-5-STATE: 2 ip sla 2 reachability Down->Up

The IOS version of the router I took these from is 151-4.M6 advanced IP services and it's a 2821 router.

I'm not sure where to look for next for troubleshooting.

Thanks!

Tate

[blau grana]

Hello,

Can you provide entire IP SLA configuration?

Do you monitor traffic load on interface? Maybe it is overloaded and icmp packets are dropped.

Best Regards

Please rate all helpful posts and close solved questions

[ap_cisco0]

Thanks for the reply...I hadn't thought of that.  I can monitor the load of the interfaces, however, it seems like when monitoring before the load was relatively light.

Here is the config for track 1 and 2:

track 1 ip sla 1 reachability

!

track 2 ip sla 2 reachability

ip sla 1

icmp-echo 206.169.109.2

frequency 10

ip sla schedule 1 life forever start-time now

ip sla 2

icmp-echo 10.10.161.2

frequency 12

timeout 7000

threshold 7000

ip sla schedule 2 life forever start-time now

Thanks,

Tate

[blau grana]

Check also if some input/output policy is not configured on interface. Maybe for ICMP is not reserved enough bandwidth.

I would suggest to try some tool to measure latency (PRTG, Smokeping) I think that this will be problem.

One last thing, cisco recomends to configure timeout > threshold.

Best Regards

Please rate all helpful posts and close solved questions

[ap_cisco0]

OK I'll get a smokeping set up and start to monitor.  I do have an output policy on the interface that track 1 is tied to but not track 2.  Track 2 is the one that seems to drop more anyways.  Is there a way to see if ICMP traffic is getting dropped?

Oh and thanks for the info on timeout > threshold.  I've always been confused about that.  I'll tweak it accordlingly.

Thanks,

Tate

[ap_cisco0]

So I've been monitoring for a week with smokeping and I don't really see any latency issues.  I did noticed, however, that both of these interfaces experiencing the issue are on the same card/wic slot.  Any chance that this could somehow be a physical issue with a bad card?  Any way to dig in and see if there was an issue on the card?

The card is a HWIC-2FE

Thanks,

Tate