Possible spanning tree issue

[Matthew Frederick]

Hello all,

I have a network of around 100 switches. I have four major distribution points that are setup like this:

Switch A > Switch B

                 Switch B > Switch C

                 Switch B > Switch D

Now Switch B is a 3750X 12 port fiber switch and the other 3 are all 4507's. Now I am wanting to add a link from Switch A to Switch C and have the topology look like this:

Switch A > Switch B

                 Switch B > Switch C

                 Switch B > Switch D

Switch A > Switch C

All of the switches have the following spanning-tree configurations on them:

spanning-tree mode pvst

spanning-tree loopguard default

spanning-tree portfast bpduguard default

spanning-tree extend system-id

The ports on Switch A and Switch C for the new link had only this on the config:

switch trunk encap dot1q

switch mode trunk

When I plugged the new link in, the entire network went down for a moment until I unplugged it.

The only logs that showed up was on Switch D and it was the following:

Dec 27 10:11:14: %HSRP-6-STATECHANGE: Vlan8 Grp 1 state Standby -> Active

Dec 27 10:11:59: %HSRP-6-STATECHANGE: Vlan8 Grp 1 state Active -> Speak

Switch A is the root for vlan11

Switch D is the root for vlan1

Switch C is the root the rest of the vlans in use (there are around 60 total)

Any suggestions would be greatly appreciated. Please let me know if any additional information is needed.

Thanks very much.

[atef yamin]

Hey Matthew,

Without show commands or running config it will be tough to say what happened. Provide the running config for the switches, or their HSRP configurations as the log is showing HSRP change.

Atef

[mahmoodmkl]

Hi,

The HSRP logs are due to the loop.

For sure there was a topology change,as suggested it will be difficult to assume what happened without the show commands and config.

[Matthew Frederick]

Well I was told that I cannot publish any specific information like the running configs, so it looks like I will probably be better off just opening a TAC case. Thanks anyway!

[Terry Cheema]

Matthew,

Before you added link from SWA TO SWC, your topology was loop free and you had only one way of reaching all the switches. But after you introduced link AtoC, there became more than one paths for the switches to reach each other. I think if spanning tree was enabled correctly on the switches, the spanning tree should have taken care of this, blocked the redundant links, for example, put one of the links between A and B in BLOCKING state for the VLANS SWC is root (depending on the port priority) and topology would be again loop free (similarly block other redundant links for VLAN1 and 11). But looks like this is not happening.

I think you did the correct thing TAC may be able to speed up the resolution by remote access etc.Please keep us posted how did you go and let us know if you have any other query.

Terry

[milan.kulik]

Hi,

I agree it seems like some STP problem.

I'd suggest to check if STP is working properly in the current LAN - do the access switches see correct core switches as roots in all VLANs, e.g.?

It might be worth to permit just several (or even one) VLANs on the new trunk and observe if that kills the whole LAN again?

HTH,

Milan

[Jeff Van Houten]

You said the network went down for a moment before you unplugged the new link. How long did you leave it plugged in? You are using pvst not rapid pvst. Maybe your spanning-tree simply didn't have time to converge before you backed out.

Sent from Cisco Technical Support iPad App