because someone turned on bridging in the host, and spanning tree is doing exactly what it should as this is a loop bridging in windows XP sometimes gives this kind of hassle, but shouldnt be on a server....
the host has the same MAC address on all ports and the switch got confused and put 1 port in err-disable (used to be common on a Sun box)
1 of the wierd mix ups that sometimes happen with link aggregation
- not setting it up on both ends, or mixing 802.1ad with cisco port aggregation....
slightly related - latest one i seem to get is Ethernet WAN links that reflect packets when there is an error (or when the transmission engineers are testing - they insist on looping circuits).
cisco routers dont seem to mind too much, but a Catalyst will err disable the port very quickly - and they dont auto recover by default....
I was really fishing to try to get the original poster to bring out that kind of issue.
I've never come across that as an issue causing err-disable. What circumstances does it happen in?
Anyway I was assuming (and looking back the assumption may have been wrong) that when the OP talked about STP blocking he was seeing his switch reporting a port in BLK mode. Reading back he may have been inferring (wrongly) that because a port was down it might be due to STP.
GigE negotiation mismatch.
Not seen that one (yet...). We have all our boxes set to recover from err-disable automatically - so much better than leaving the port down.