STP failures

dsandre-toh · ‎11-02-2007

Hi,

have some questions about a recent STP issue in our network. After reviewing several Cisco STP design & troubleshooting guides, the documents repeatedly indicate that a STP failure is usually the result of a blocked port transitioning to forwarding mode, thereby causing a loop.

- is this truly the cause of most STP failure issues?

- if so, would it make sense to determine where the blocked ports are in the network and focus our attention on those ports for:

duplex mismatch,

udld,

packet corruption,

resource errors,

portfast config errors,

stp parameter tuning, as per the design guides ??

- or, is there another major cause of STP failure we should be looking at?

thanks,

DS

xzevallos · ‎11-02-2007

In my experience, most of these issues have been caused by the following:

1. A switch with two uplinks toward the root, where one is normally blocked, and the other link that is normally the designated port is either flapping or has many CRC errors. The flaps on the designated port cause the normally blocked port to unblock and block again continuously as a result of STP - this causes temporary loops when it is unblocked and the designated port comes up again. A high number of CRC errors causes alot of missed BPDUs received on the designated port, which has the same effect as a flapping port.

2. With fiber connections, a uni-directional link condition causes bridging loops to occur if UDLD is not enabled on both ends, becaue STP is unable to detect this condition.

Problems like these can be alleviated by using UDLD on all ports, and Uplink Fast if STP is used, or Rapid-STP instead.