C3560X HSRP problem

[inb]

Top-C3560X-48P-E 2 units:
L3 mode, PVST, Downlink port Trunk, HSRP(vlan 10, 20, 30 ...)
OS version-15.2(4)E9

Sub-C9300-48P-A several groups:
L2 mode, RSTP, Uplink port TRUNK
OS version-16.12.4

L3#1
int vlan 10
ip address 192.168.109.2 255.255.255.0
standby 10 ip 192.168.109.1
standby 10 priority 105
standby 10 timers 3 4
standby 10 preempt

L3#2
int vlan 10
ip address 192.168.109.3 255.255.255.0
standby 10 ip 192.168.109.1
standby 10 timers 3 4

HSRP is operating with the above settings.

There is no work or topology change,
Often in L3 #2, HSRP goes into Active state and then goes into Standby state.
Attached PNG file (analyze HSRP packet through wireshark)

Normally, a problem occurs after sending and receiving hello packets every 3 seconds normally (all SVIs are the same).
No. In 120332, 120343, L3 #2 declares Active.
No. At 120359, L3 #1 sends a Coup packet to declare Active
No. At 120510, L3 #2 switches to Speak.

The above process (No. 120332~120510) took place within 1 second when looking at the time,
This causes the switch's HSRP Active/Standby to be switched.
Connected clients suffer ping loss.

Usually, hello timer and hold timer should be three times different.
However, we are using the service by adjusting hello 3 and hold 4 respectively.

In other buildings L3 #1,2 there is no problem with the same configuration.

Which part should we check?
Please help.

[balaji.bandi · Accepted solution]

can you post below information :

show standby brief

what version of hsrp ?

do you have spanning tree configure  ? what is the status of that VLANs?

wht other Logs you see on both the device when they go up and down ?

here is some tips for troubleshoot :

https://www.cisco.com/c/en/us/support/docs/ip/hot-standby-router-protocol-hsrp/10583-62.html

[inb]

Thank you for answer.

show standby brief-> Active/Standby is changed from L3 #1,2.
Normally, L3 #1 has Active on all VLANs.

what version of hsrp?

-> hsrp v1

do you have spanning tree configure?

-> L3 #1 PVST, spanning-tree vlan x,x,x,x priority 4096, backbonefast
L3 #2 PVST, spanning-tree vlan x,x,x,x priority 8192, backbonefast
L2 #1,2,3... RSTP, L3 #2 uplink-cost 100,000(block)

what is the status of that VLANs?
-> SVI protocol and link up are maintained.

what other Logs you see on both the device when they go up and down?
-> There is no up/down log.
L3 #1: Switch to Act->Std
L3 #2: Switch to Std->Act
As soon as the state transitions, it behaves just like below
L3 #1: Switch to Std->Act
L3 #2: Act -> Switch to Std

Thank

and reguards.

[balaji.bandi]

First suggestion is move immediate to HSRP v2 and monitor before we do anything else. (like spanningtree)

[inb]

Thank you for answer.
We can't switch to hsrp v2.
Only when the exact cause is identified can proceed.
I will refer to the guide document you gave me.
Thank you for your support.

[paul driver]

Hello
Either way you say those timers are "working" for other devices they are still aggressive regards the 1 second interval between sending the hello and the hold timer threshold, what would require such a short hold interval?

Are you pushing large amounts of traffic?
Whata the cpu utilization of those switches?

Have you:

• Tried increasing the hold interval, does this rectify the HSRP transition?
• Checked the  physical connections between these two switches?

[inb]

Thank you for answer.
The hello/hold timer value is a result that has been verified to have a minimum switching time through numerous tests in advance.
All other building backbones are in operation with the same hello 3sec and hold 4sec configuration.
Traffic is in use around 1/10~3/10, and there is no problem with interlink and STP configuration between devices.
There are no physical problems either.
No OS bugs have been identified.
Thank you for your support.