HSRP flap caused by IP SLA flaps

[imran.exte]

Hi All,

We are facing a concern of IP SLA flaps which is further causing HSRP to flap, every min there is log of the switch.  However the interesting traffic which traverse the SVI interface (ip sla is tracked) is Real time in behavior and applications would not expect any kind of delays. Therefore, I seek to take an expert advise as the following configuration posted below requires any fine tuning of the parameters. Please share your valuable thoughts as there is no impact to operations but cant wait to have one.

ip sla 1
icmp-echo x.x.x.x
threshold 500
timeout 1000
frequency 3
ip sla schedule 1 life forever start-time now

track 1 ip sla 1 reachability

Thank you.

[Jaderson Pessoa]

Hello

I suggest to you input this parameter

track 1 ip sla 1 reachability
delay up 10 down 10 (seconds)

[Georg Pauwen]

Hello,

what are you tracking, the HSRP group ? And what is the target of your icmp echo ? Post your HSRP configs as well as a schematic drawing of your topology...

[imran.exte]

Thank you for the responses to Jaderson & Georg.

Hi Georg,

We have in our scenario, a high BW link (active) and low BW link (backup) on each WAN router A and B. HSRP is configured on switches A & B which are connected to corresponding rtr A and B. So we always want the High BW link to be preferred. Hope this throws some light on the topic, however if there is still a schematic representation desired I can provide one by tomorrow.

Please find the below config,

ip sla 1
icmp-echo x.x.x.x (remote end IP of High BW interface)
threshold 500
timeout 1000
frequency 3
ip sla schedule 1 life forever start-time now
!
track 1 ip sla 1 reachability
!
interface Vlan45

ip address y.y.y.y 255.255.255.248
ip pim sparse-dense-mode
standby 45 ip y.y.y.z
standby 45 priority 110
standby 45 preempt
standby 45 name ABCD
standby 45 track 1 decrement 20

[Georg Pauwen]

Hello,

post the output of:

show ip sla 1 statistics

Right now you are sending a ping every 3 seconds. Jaderson Passoa made a very good suggestion, which is to increase the down delay (effectively telling the SLA to be considered down after 3-4 missed pings). That will probably get rid of the flaps.

Or send a ping every let's say 30 seconds and change the down delay to 40, which means the ping has to fail twice before the SLA is considered down.

ip sla 1
icmp-echo x.x.x.x (remote end IP of High BW interface)
threshold 500
timeout 1000
frequency 30
ip sla schedule 1 life forever start-time now

!

track 1 ip sla 1 reachability
delay down 40 up 10

[Jaderson Pessoa]

yes.. i think that our suggestion to change delay will solve his problem :)

[imran.exte]

Hi,

Please find the below output of Show IP SLA Statistics;

IPSLA operation id: 1
Latest RTT: 3 milliseconds
Latest operation start time: 07:30:55 UTC Sat Nov 2 2019
Latest operation return code: OK
Number of successes: 642
Number of failures: 44
Operation time to live: Forever

Thank you for the suggestions.

I would like to make you guys aware that the application data that is forwarded on this SVI interface is highly real time traffic in nature and will the above proposed solution of 30 sec be able to take of care of the same or there is another fine tuning mechanism with less frequency rate like 3, only based on nature of RT data will it fail-over seamlessly without any unexpected flaps.