HSRP - What causes my Active/Standby routers swap?

[Ellerby01]

I am a litte confused in my understanding of what will cause HSRP connections to fail over, could someone please offer some clarification regarding a HSRP configuration I have?

I am using two Catalyst 3750's in a HSRP configuration.

I have multiple networks connected to these 'Routers' with each network having one connection to each of the routers. Each pair of connections for a network is then configured as a standby group to provide HSRP for that network.

If I lost one of the connections that participate in one particular standby group, will the active standby connections swap to maintain connectivity via the other router? Or does a physical 3750 have to fail before these connections swap?

There is no interface tracking configured on these devices.

Apologies if I'm asking the obvious, but when reading Cisco's literature on this, the the terminology refers to 'Routers' swapping. From the perspective of my system does this actually mean 'the router ports for this particular network'

I hope I've made sense!

Thanks fror any help you can offer

[John Blakley · Accepted solution]

Hi Matthew,

HSRP uses a multicast address (224.0.0.2 for version 1 and 224.0.0.102 for version 2) to find other hsrp members. The source address that it uses is the real ip on the interface when sending these packets. When it finds those members, it starts sending unicast hello packets (Hellos come from both directions). If that link goes down, the standby will not get a hello from it's old active so it assumes the role.

That being said, if you have a vlan that has a single member and that member has a connection to each 3750 and the link fails on the primary, the vlan will go down. With the vlan going down (because it no longer has members), the other 3750 will take over the role because the old active is no longer sending hellos to. Your interface does have to stop responding in some way if you're not using tracking though. Take the previous example and say that you have 5 members on this vlan. If one of those members go down (server dies), that's not enough to have the standby take over because the vlan itself didn't go down. You'd have to lose all members in order for it to go down.

HTH,
John

*** Please rate all useful posts ***

[Ellerby01]

OK thanks,

So once discovered by multicast, a standby group's member interfaces will unicast hello packets to each other. If the active interface goes down, the hello packet will not be recieved by the standby and so the standby interface will take over.

I assume other standby groups on these 3750's that don't have any faults are unaffected by one particular standby' s members swapping?

Thanks for your help.

[Richard Burts · Accepted solution]

Matthew

Yes if one standby group has a failure event that causes the backup to take over and become primary it would not affect other standby groups on that switch/router.

HTH

Rick

[Ellerby01]

Thanks for your help