Loopback detected error on 2960 switch

Unanswered Question
Jan 17th, 2008
User Badges:

I have 2960 switch that is connected to 2 3560 switches for redundancy through the GE ports on the 2960. 2 days ago, I started getting loopback detected errors which disabled the ports and effectly took down the switch. I got the following response from the output interpreter when I plugged in my log messages:


%ETHCNTR-3-LOOP_BACK_DETECTED: Interface [chars]: [chars]

Explanation: The problem occurs because the keepalive packet is looped back to

the port that sent the keepalive. Keepalives are sent on the Catalyst switches

in order to prevent loops in the network. Keepalives are enabled by default on

all interfaces. You see this problem on the device that detects and breaks the

loop, but not on the device that causes the loop.


My question is this. Is the loopback caused by the fact that the switch is connected to 2 different switches (I thought that's what STP was for) or is this just between the two individual ports on the two switches?


Thanks,

Greg

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (1 ratings)
Loading.
Kevin Dorrell Thu, 01/17/2008 - 07:25
User Badges:
  • Green, 3000 points or more

You are right that normally STP will protect you from this sort of thing. But clearly you have a loop somewhere. I presume up to 2 days ago everything was working fine, even with both uplinks.


Do you have bpdufiter enabled on any of your access ports? If you do, then my advice is don't! That will disable the protection that STP gives you.


Another precaution you could take is to enable bdpuguard on all access ports. This will disable the port if any user decides to do something malicious with a cross-cable.


The other thing to look out for is unmanaged switches on your network of the type you buy in the supermarket. They should be avoided at all costs. Some users seem to have an irresistible urge to connect them to two access ports at once.


Finally, consider whether you have any XP PCs with two NIC cards. If both are plugged into the same VLAN, they will bridge, and as far as I know they won't run Spanning Tree.


Are you 3560 switches giving you any messages about addresses flapping between two ports?


Kevin Dorrell

Luxembourg


justin.wade Thu, 01/17/2008 - 07:31
User Badges:

Greg,


To answer your question this is not a spanning tree issue. Spanning tree is used to eliminate bridging loops at layer 2. You are somehow seeing your keep alives returning on the same interface on which they were initially sent.


This sounds like it is either a hardware issue or cabling problem. I am guessing these switches are connected via copper? If so I would check my terminations, maybe there somehow shorted to one another, switches use pins 1,2 3,6. So if pin 1 was shorted to pin 3 and 2 was shorted to 6, you would see this issue.


Hope this helps.




Actions

This Discussion