We recently migrated our server farm to a set of 6500/Sup2 running IOS 12.1(13)E6.
The boxes are configured with Rapid-PVST, and all server access ports are configured with PortFast, thus acting as RSTP Edge ports:
r-edb1-14/sa-1#sh run int fast 9/23
Building configuration...
Current configuration : 188 bytes
!
interface FastEthernet9/23
description Digizuite01-A
no ip address
duplex full
speed 100
switchport
switchport access vlan 264
switchport mode access
spanning-tree portfast
end
r-edb1-14/sa-1#sh spanning-tree int fast 9/23
Vlan Role Sts Cost Prio.Nbr Type
---------------- ---- --- --------- -------- --------------------------------
VLAN0264 Desg FWD 19 128.535 Edge P2p
http://www.cisco.com/warp/public/473/146.html clearly states that only NON-Edge ports should cause a Topology Change.
But we're logging concurrent SNMP traps of Topology Changes and Link Up/Down on server access ports, indicating that the switch treats these Link Up/Down events as Topology Changes (9/23 and 9/38 on the two switches are two NICs on the same server, which explains why they're in sync):
1057729771 7 Wed Jul 09 07:49:31 2003 r-edb1-14.dr.dk - linkUp trap received from enterprise cisco.1.283 with 4 arguments: ifIndex=131; ifDescr=FastEthernet9/23; ifType=ethernetCsmacd; locIfReason=up;1 .1.3.6.1.6.3.1.1.5.4.1.3.6.1.4.1.9.1.283 0
1057729771 7 Wed Jul 09 07:49:31 2003 r-edb1-15.dr.dk - linkUp trap received from enterprise cisco.1.283 with 4 arguments: ifIndex=98; ifDescr=FastEthernet9/38; ifType=ethernetCsmacd; locIfReason=up;1 .1.3.6.1.6.3.1.1.5.4.1.3.6.1.4.1.9.1.283 0
1057729771 4 Wed Jul 09 07:49:31 2003 r-edb1-14.dr.dk - Spanning tree topology change from enterprise dot1dBridge;2 .1.3.6.1.2.1.17.0.2 0
1057729771 4 Wed Jul 09 07:49:31 2003 r-edb1-15.dr.dk - Spanning tree topology change from enterprise dot1dBridge;2 .1.3.6.1.2.1.17.0.2 0
1057729777 4 Wed Jul 09 07:49:37 2003 r-edb1-15.dr.dk - Spanning tree topology change from enterprise dot1dBridge;2 .1.3.6.1.2.1.17.0.2 0
1057729777 4 Wed Jul 09 07:49:37 2003 r-edb1-14.dr.dk - Spanning tree topology change from enterprise dot1dBridge;2 .1.3.6.1.2.1.17.0.2 0
1057729777 7 Wed Jul 09 07:49:37 2003 r-edb1-15.dr.dk - linkDown trap received from enterprise cisco.1.283 with 4 arguments: ifIndex=98; ifDescr=FastEthernet9/38; ifType=ethernetCsmacd; locIfReason=Lost Carrier;3 .1.3.6.1.6.3.1.1.5.3.1.3.6.1.4.1.9.1.283 0
1057729777 7 Wed Jul 09 07:49:37 2003 r-edb1-14.dr.dk - linkDown trap received from enterprise cisco.1.283 with 4 arguments: ifIndex=131; ifDescr=FastEthernet9/23; ifType=ethernetCsmacd; locIfReason=Lost Carrier;3 .1.3.6.1.6.3.1.1.5.3.1.3.6.1.4.1.9.1.283 0
1057729779 7 Wed Jul 09 07:49:39 2003 r-edb1-15.dr.dk - linkUp trap received from enterprise cisco.1.283 with 4 arguments: ifIndex=98; ifDescr=FastEthernet9/38; ifType=ethernetCsmacd; locIfReason=up;1 .1.3.6.1.6.3.1.1.5.4.1.3.6.1.4.1.9.1.283 0
1057729779 7 Wed Jul 09 07:49:39 2003 r-edb1-14.dr.dk - linkUp trap received from enterprise cisco.1.283 with 4 arguments: ifIndex=131; ifDescr=FastEthernet9/23; ifType=ethernetCsmacd; locIfReason=up;1 .1.3.6.1.6.3.1.1.5.4.1.3.6.1.4.1.9.1.283 0
1057729779 4 Wed Jul 09 07:49:39 2003 r-edb1-15.dr.dk - Spanning tree topology change from enterprise dot1dBridge;2 .1.3.6.1.2.1.17.0.2 0
1057729779 4 Wed Jul 09 07:49:39 2003 r-edb1-14.dr.dk - Spanning tree topology change from enterprise dot1dBridge;2 .1.3.6.1.2.1.17.0.2 0
Am I overlooking something here or just facing a bug?
My obvious concern is that we end up facing unicast flooding, as Topology Changes will cause a flush of MAC address tables in the switches in the server farm.
