cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
2474
Views
0
Helpful
10
Replies

ASR9k strange bvi interface behaivor

Hi.

We are running asr9006:s as backbone nodes. We have mstag accessrings wich connects different acces switches through out the network. We have been running this for a long time (3 years). After we upgraded from 4.1.0 to 4.3.3 we are experience som strange behavior on some of the vlans running accross the mstag.

Let me try to explane..

One of the vlan is connected to 2 asr9k in a accessring. The particular vlan is configured betweeen the two 9k:s with bridge grops and bridge- domains according to our standard setup for the backbone. Both routers have a configured bvi for this bridge-domain, both configured equally, but one of them is in shutdown mode.

Suddenly we are experiencing packet drops and we are unable to ping end nodes in this specific vlan wich is routed at the bvi The problem is intermittent and seems to last for a couple of minutes before it starts to work again.

While troubleshooting this problem i noticed that if we disabled the bvi on one router and enabled it on the other, the problems disappear.  Changing back (activating the bvi on the first 9k), makes the problem come back again after a while.

All pseudowires and mstag check commands indicates that the bridge-domain and access ring is ok. Since we have been running this setup for a while i am beginning to think this is a 4.3.3 bug. But why does the other router handle the situation?

 

Any ideas?

 

 

1 Accepted Solution

Accepted Solutions

hi andreas,

this situation will require some troubleshooting and some deeper investigation. It might be best to open a TAC case for this so we can do a screen share and verify the various points in the topology to see where the packet drops occur potentially.

One thing you could check is to see if there are STP convergence events happening that potentially, temporarily, bring down a link or put it in blocking instead of forwarding.

The other thing to verify is seeing if there are paritcular NP counters that reflect a rate that seems to be related to the drops.

The ping loss to the BVI can either be caused by drops in hardware, lpts (control plane policing) or at the software level.

There are too many dependencies here to give a solid confined answer for which that tshoot session would be necessary.

regards

xander

View solution in original post

10 Replies 10

STEVEN RUBIN
Level 1
Level 1

I have a case from April that had a similar behavior.   4.3.4 and 5.1.2 solved it for us.  The 3 ddts's on my case are CSCuo36365,CSCui95200,CSCui53902.   The workaround was static mac entries.

 

Thanx Steve.

I now realize i have made a typo in the above text. We did infact upgrade to 4.3.4 and NOT 4.3.3, my bad.

 

I assume none of the 3 bug id exist in 4.3.4?

 

/Andreas

 

Actually reviewing my notes, we went straight to 5.1.2, so I never did verify it was fixed in 4.3.4.   I seem to remember this problem being limited to Trident cards only.  Is that what you are using?

hey guys,

CSCuo36365 is dupe of CSCui95200 which is fixed in 434 onward.

CSCui53902 is fixed in 434 onward also.

the bvi like an svi is brought down when there are no active circuits in the BD, that could be part of the issue?

the interface manager trace may hold some clue describing what the reason was for bringing and keeping down the bvi potentially.

it is hard to say what is going on without seeing those debugs and details, and a confgi would help also. one other reason for bringing down an interface is a duplicate ip address.

regards

xander

Xander,

It doesn't bring down the interface, just dropping packets somewhere. Pinging the cpe ip-address from the bvi works intermittent.

 

 

//Andreas

hi andreas,

this situation will require some troubleshooting and some deeper investigation. It might be best to open a TAC case for this so we can do a screen share and verify the various points in the topology to see where the packet drops occur potentially.

One thing you could check is to see if there are STP convergence events happening that potentially, temporarily, bring down a link or put it in blocking instead of forwarding.

The other thing to verify is seeing if there are paritcular NP counters that reflect a rate that seems to be related to the drops.

The ping loss to the BVI can either be caused by drops in hardware, lpts (control plane policing) or at the software level.

There are too many dependencies here to give a solid confined answer for which that tshoot session would be necessary.

regards

xander

Hi.

 

Yes i have filed a TAC case on this.

 

Thnx

/Andreas

ok let me see if I can find it and give the tac eng some pointers if necessary.

xander

Hi,

 

Do you have netflow configured on the BVIs?

Hi.

 

Nope, there is no netflow configured on the BVIs.

 

//Andreas
 

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: