Port moving from forwarding to .... forwarding?

[johnnylingo]

After setting "logging event status" on my access switches, I see the following message every few hours for several of the ports simultaneously:

%SPANTREE-6-PORT_STATE: Port Gi0/X instance XYZ moving from forwarding to forwarding

Where XYZ is the STP instance (this switch is running RPVST).

I'm a little confused how exactly a port changes from forwarding to forwarding. Does this message indicate there's a Spanning Tree topology change and the tree is being reset?

[naveen_b81]

You might have configure portfast on these ports and there might be a topology change happening if you have not configured these ports in host mode, when any system/access port comes up.

[johnnylingo]

The ports going from forwarding to forwarding are configured with portfast.

I turned on STP debugging and indeed the access switch is seeing a loss of BPDUs from the distribution switch on certain VLANs. This appears to then cause the STP instance to basically reset.

switch#sh debug

Spanning Tree:

Spanning Tree event debugging is on

Spanning Tree root changes debugging is on

switch#

Sep 24 17:10:41: RSTP(123): Po1 rcvd info expired

Sep 24 17:10:41: RSTP(123): updt roles, information on root port Po1 expired

Sep 24 17:10:41: RSTP(123): Po2 is now root port

Sep 24 17:10:41: RSTP(123): Po1 blocked by re-root

Sep 24 17:10:41 UTC: %SPANTREE-6-PORT_STATE: Port Gi0/11 instance XYZ moving from forwarding to forwarding

Sep 24 17:10:41 UTC: %SPANTREE-6-PORT_STATE: Port Gi0/12 instance 123 moving from forwarding to forwarding

Sep 24 17:10:41: RSTP(123): Po1 is now designated

Sep 24 17:10:41: RSTP(456): Po2 rcvd info expired

Sep 24 17:10:41: RSTP(456): Po2 is now designated

Sep 24 17:10:42: RSTP(456): updt roles, superior bpdu on Po2 (synced=0)

Sep 24 17:10:42: RSTP(456): Po2 is now alternate

Sep 24 17:10:42: RSTP(123): updt roles, superior bpdu on Po1 (synced=0)

Sep 24 17:10:42: RSTP(123): Po1 is now root port

Sep 24 17:10:42: RSTP(123): Po2 blocked by re-root

Sep 24 17:10:42 UTC: %SPANTREE-6-PORT_STATE: Port Gi0/11 instance 123 moving from forwarding to forwarding

Sep 24 17:10:42 UTC: %SPANTREE-6-PORT_STATE: Port Gi0/12 instance 123 moving from forwarding to forwarding

Sep 24 17:10:42: RSTP(123): Po2 is now alternate

Sep 24 17:10:43: RSTP(123): Po1 received a tc ack

The question is why is the access switch no longer seeing BPDUs from the distribution. I don't see any errors on the ports.

[johnnylingo]

OK, so when I converted the 6500s from PVST to Rapid-PVST, I got the following warning which shed some light on the subject:

%PM-SP-4-LIMITS: The number of vlan-port instances on module X

exceeded the recommended limit of 1800

The module in this case is a WS-X6548-GE-TX, so not too old. However, we have 75 VLANs configured on 35 ports, which works out to 2,600 VLAN port instances. So that would explain why the BPDUs were timing out.

I setup VLAN filtering on the trunk ports and that seems to have resolved the issue. Converting to MST would also fix it.

Cisco TAC also recommended increasing max-age timer on the STP root, which lessened the frequency but didn't make the issue totally go away.