HSRP problem on cisco 2611: "lost" ip address

[eabbink]

Hi,

I'm having a little annoying problem with a 2611. The 2611 is deployed as follows:

both ethernet interfaces go into the same switchstack and belong to bridge group 1 (interface bvi1) which is set to do ip routing.

The bvi1 interface has a configured ip address of 10.10.10.11/24 and has a HSRP address of 10.10.10.10. The backup router would be 10.10.10.12, except it's currently down due to a malfunction

(relevant configuration commands are included below.)


when the bvi1 interface comes up everything works fine and all servers etc are reachable. After some amount of time
however the 10.10.10.10 ip address is no longer reachable and the servers behind the router arent reachable also. The 10.10.10.11 ip address keeps working fine and can also be used to route to the servers.

According to the router it's at this time still in "possession" of the 10.10.10.10 ip address and HSRP "active". (and it really cant be in standby mode, since there is no other router currently)

A manual shutdown/no shutdown of the bvi1 interface will fix the problem (temporary). As such the HSRP config is obviously not really helping us...

Does anyone have an idea whats going wrong here?

thanks,

Esger



details:

Cisco 2611 router
IOS 12.0(7)T

config:

bridge irb
bridge 1 protocol ieee
bridge 1 route ip

ethernet0/0
no ip address
bridge-group 1
no ip mroute-cache
no ip directed-broadcast
no cdp enable

ethernet0/1
no ip address
bridge-group 1
no ip mroute-cache
no ip directed-broadcast
no cdp enable

bvi1
ip address 10.10.10.11 255.255.255.0
no ip redirects
no ip directed-broadcast
ip nat inside
standby priority 120 preempt
standby ip 10.10.10.10

 

[eabbink]

anyone please?

 

[gconnect]

have you attempted to log or analyzed the output from the 'debug standby' cmd?

is the bvi and relevant ethernet interfaces in the up/up states when this occurs? 'show interfaces'

 

[eabbink]

I havent got a debug from during the actual event of losing the ip. I do have checked the standby status after loss and the router still owns the ip and is in active state as far as it is concerned.

the bvi interface and the ethernet interfaces are all fine. this is also proved by the fact that the other ip address on the bvi interface never has any problems.

I also know for certain that the loss of ip is not temporary, but remains permanent.

Esger

 

[gturrubi]

I believe in the bvi interface you need to state the mac address of the ethernet interface that has ip address 10.10.10.10.

 

[eabbink]

wouldnt the router take care of that itself? As I understand it the hsrp config will cause the router to create a virtual mac address and use that instead of the one from the interface.

so wouldnt an extra mac for the bvi just simply be ignored?

and remember, it works right when freshly booted. it's only after days (or even weeks) that the problem occurs.
after your remark about the mac address I started thinking it could maybe have somthing to do with the bvi interface switching from active ethernet interface. In this case the switches (two trunked 2950T's) could possibly have a problem since then the virtual mac address would have hopped ports.

But I'm pretty certain once it's stopped working it's permanent (or at least hours in duration) and wouldnt the hopping cause just a fairly temporary problem?

Esger

 

[oj88]

I may be mistaken but, as I understand it, there could be an STP issue here. As you've bridged both ethernet interfaces on the router (and eventually enabled STP), this would appear to the outside world as a swith. And the fact that you've connected another switch (2950) in parallel, you'd probably have some ports forwarding and one in the blocking state (either on the switch or on the router).

What I'd do on is to enable FEC on both the router (port-channel and channel-group) and the switch if I want a higher aggregate bandwidth and some degree of link redundancy. However, this only works on Fastethernet interfaces and up.

Orlando Palomar Jr
CCIE# 11206, CCNP
CIPT Operations Specialist
Phil-Data Business Systems, Inc.

 

[fmonteiro]

I agree with oj88 that we could have a seriously bridge loop here. I suggest trying the following command, on both routers:
no bridge 1 bridge ip

 

[eabbink]

Unfortunately it's just a 2611, meaning no fastethernets..

I just tried the 'no bridge 1 bridge ip' command but that results in no communication possible at all. (Please note that I only did some quick testing as it's in a live network)


On a sidenote, up until recently only the HSRP ip was failing us but we've now experienced at least one instance where the normal address lost connectivity. On all occasions it seems to be just 1 ip address that looses connectivity though.

In both cases a shutdown/no shutdown of the bvi1 interface immediately fixes the problem.


thanks for the help.

Esger