cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
444
Views
0
Helpful
5
Replies

Applications losing connections since switch upgrade.

dustinn3
Level 1
Level 1

We recently replaced all of our switch infastructure. We had a mixture of Dell's, 3Com's, and Cisco's before. We have 2 3750g's for our core switches and 2960's for our edge switches. The whole network is gigabit now and our utilization is really low, it doesn't even show up on the traffic graphs. However at random times throughout the day we have clients in seperate building complaining that they are all losing connections to applications on seperate servers. I have checked all the trunk ports, server ports, and client ports and there are zero errors. I can ping the servers with no dropped packets. The problem started immediately after installing the new switches and since it's on multiple servers at the same time in multiple locations it seems like it has to be a network problem.

Any ideas?

5 Replies 5

Joseph W. Doherty
Hall of Fame
Hall of Fame

Spanning tree issue? What's the topology look like? Running any spanning tree?

You are correct it is a spanning tree issue. We don't have redundant links on any of our switches so spanning tree shouldn't be blocking. We have a core stack of 3750g's with trunks to 2960's and a couple of 3750's. Most of the switches are daisy chained from floor to floor.

I have finally figured out that several times a day I am getting the following error on several switches.

%SPANTREE-2-LOOPGUARD_BLOCK: Loop guard blocking port GigabitEt

hernet0/? on VLAN0001.

%SPANTREE-2-LOOPGUARD_UNBLOCK: Loop guard unblocking port Gigab

itEthernet0/? on VLAN0001.

It is happening on 4 out of 26 switches at the same time. In different locations. 2 are 2960's and 2 are 3750g's. It happens on the trunk port on all 4. Some are connected back to a core stack of 3750's and some are daisy chained. However, the switches in between aren't getting the error. I thought at first it could be a bad cable, but not on 4 switches. Any ideas?

I think I may have found part of the problem. A proxim wireless bridge on the other side of a wireless link was acting as the root bridge. I'm assuming that if it ever loses connection it's making the others shut the ports down. However, I don't understand why only 4 of our new switches are actually blocking ports.

Thanks,

If you don't have any redundant paths, you could try insuring spanning tree is deactivated on all your equipment. (One guess for cause is "confusion" between mixed vendor equipment when you don't really intend to use STP.)

I'd still like to leave it on, just in case someone decides to make a loop by accident.

I finally found the problem by examining all the running configs. The switches going down had a cisco-global macro on that enabled loopguard default on all ports, where the ones not dropping didn't. I just removed the command on them for now, which should stop it from disabling the ports when it doesn't hear from the root. However, since I made the core the root, I wouldn't think it would be a problem anymore either.

Thanks,

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:

Innovations in Cisco Full Stack Observability - A new webinar from Cisco