Campus-Latency-STP/CEF

[vyadav1980]

hi guys

now i have a campus network with a core location and one zonal location with a 10G backbone.its a star-bus topo.

now my issue is if i ping a network switch which is "directly attached" to my core location and my pc is hooked to the core switch 4500-SUPV-10G, i get a response of 500ms, cant belive it , the network is slow, departments are complaining.

i disabled spanning tree first as i was getting topology change from various ports.

then i get :

Error Message C4K_EBM-4-HOSTFLAPPING:Host [mac-addr] in vlan [dec]

is flapping between port [char] and port [char]

Explanation: The specified host is detected as a source address on multiple

ports. Typically, a host is supposed to be learned on only one port. A spanning

tree loop is the most common cause of this condition.

so disabled stp.but no change..

for soem minutes it comes to 1ms then goes back to 500ms..in the morning and evening its 1 ms but throughout the day its...500-600 ms..

i thought it could be the switching method, as im on cef and fast switching.

im giving soem configs and input.

Im Stuck!!!!

i have switches of numb 90, i can ping all of them, but will not show in cef..i know that its not doing the inter vlan routing..it purely layer2..

just cant fig it out.

Shukky

India

[mnewnam05]

Do you have hub somewhere?

[Francois Tallet]

Hi Shukky,

Bridges are never slow;-) The 500 ms you delay you are seeing is definitely not because your traffic gets stuck in some queues in the network. Believe me that no Cisco switch (or third party ones) have enough buffer to queue 500ms of traffic at 10Gig.

My guess is that you are indeed experiencing a bridging loop or at least high link utilization in your network. The delay you are mentioning could come from the station you are pinging, which is probably hit by lots of traffic and which is slow to answer.

First, disabling STP is a very bad move. The warning message you saw is indeed typical of a bridging loop, which is the worst case scenario in a L2 network.

Re-enable STP and please, post a diagram of your network. Considering the warning message you have, you certainly have some redundancy in your network (some redundancy you may not expect btw). So first discover what your network is exactly. Then, if it's an emergency and your network is really unusable, disable the redundant links manually. If you have time to troubleshoot, then look for the STP state and role on an affected vlan along the physical loop that you have identified and let us know.

Good luck!

Francois

[vyadav1980]

hi francois

thnx a ton for ur valuable input, i always knew it was the stp causing, ok now i did what my common sense said, i re-enabled stp and started working towards the stp root bridge id which was found to be some switch which we didnt even installed. we r doing part new and part upgradation for the campus so these stray cases of soemone puting a hub or switch is there.

my topology is simple..1 core 4510R with star edge catalysts, a 10G backbone to Cat4503R as zonal, again a star topo there.

but now its working fine after applying portfast and bpduguard on all major siwthces giving these messages, after these 2 configs, the network stabilised.

i have stacks of 8 switches and 7 switches on soem locations, do i need to worry about the stp network Diameter , as between the 2 remote swicthes there r 10 switches in between.

heres my topology.

cheers

Shukky

India

[Francois Tallet]

Hi Shukky,

Your topology suggests indeed that STP should not do much. Now, practically, you had the symptoms of a bridging loop so indeed, it is possible that you got some undocumented links and configuration from the existing network.

BPDUguard is a feature that will bring down a port should it receive a BPDU. If this configuration stabilized the network, did you see any port being shut down? That would be a clue that a bridge is connected to a port that you expected to be an access port.

Regards,

Francois