faulty switch that brings the whole network down

[rob_xx1717]

I have a network consisting of a number of the cisco switches (all the switches are the same). I'm using all static IP addresses. There are no routers just C2950G switches. The network works fine. All the switches are able to communicate/(ping) with each other. I have no address conflict. But I also have one, presumably faulty switch (same model as the rest). Once I put this faulty switch on the network all the communication between the rest of the switches dies. When I used the faulty switch alone with other devices (say workstations) it worked fine. Only with connection with other switches the faulty switch brings the whole network down. I configured all the switches the same (static IP addresses, passwords, etc.). Is there any means for me to isolate (ARP, STP, SNMP or anything else) and correct the problem. I'm thinking that maybe the faulty switch is broadcasting some sort of signal that disables the communication between all the other switches. At this point I'm leaning towards a hardware problem. I appreciate you help with this issue.

rob

[bbanis2k]

Leave the switch disconnected from the network and connect to it via a console cable.

After doing that please paste the config here.

BBanis2k

[Georg Pauwen]

Hello Rob,

chances are that the faulty switch has the highest VTP revision number and takes over as VTP server for your entire domain. If that happens, only the VLANs configured on that switch are propagated throughout the domain, everything else on all other switches will be deleted.

A workaround for this is to configure the switch for a non-existent VTP domain and then back to the original one, this will reset the revision number. Or set the switch to transparent and then to client mode.

Does that make sense ?

Regards,

GP

[bbanis2k]

Below is an example of the option in global config mode.

(config)#vtp mode transparent

[robphill]

Do a 'show post' and if you get any POST Failures you probably have some sort of hardware problem. I have seen where a bad transceiver or switchport floods traffic on the network and slow it down. Never seen it kill a network completely however. Worth looking at though. If you have POST Failures the switch needs to be replaced whether it is causing your issue or not because there is a hardware fault somewhere in the box.

I'd also look at VTP and see if it is in server mode with a higher revision number than the other switches.