Odd traceroute information / potential routing issue.

[jeff.vargas]

I've got a single core 6500 switch attached to two 4500 distribution switches and the 4500 switches have a port channel between them. Basically a triangle. When I trace to a client that is connected to one of the 4500's from the core 6500 I get the following:

Type escape sequence to abort.

Tracing the route to 10.227.33.13

1 10.227.1.122 0 msec

10.227.1.118 4 msec

10.227.1.122 0 msec

2 10.227.33.13 0 msec 0 msec 4 msec

The 10.227.1.122 and the 10.227.1.118 are the 4500's. I should only get the 10.227.1.122 reponse on the path to the 10.227.33.13 client. It should look like this:

Core1#trace 10.227.33.13

Type escape sequence to abort.

Tracing the route to 10.227.33.13

1 10.227.1.122 0 msec

2 10.227.33.13 0 msec 0 msec 4 msec

Has anyone seen this before in a switched environment? I'm banging my head and I can't figure this out. Please help.

[Georg Pauwen]

Hello Jeff,

is the 6500 switch the root for your VLANs ?

Regards,

GP

[jeff.vargas]

Yes, the 6500 is the root:

C6509#sh spanning-tree detail

Bridge group 9 is executing the dec compatible Spanning Tree protocol

Bridge Identifier has priority 128, address 00d0.97e4.1800

Configured hello time 1, max age 15, forward delay 30

We are the root of the spanning tree

Topology change flag not set, detected flag not set

Number of topology changes 2 last change occurred 1y36w ago

from Vlan20

Times: hold 1, topology change 45, notification 1

hello 1, max age 15, forward delay 30

Timers: hello 0, topology change 0, notification 0, aging 300

[sundar.palaniappan]

Show ip route 10.227.33.13 on your 6500 should have 2 routes pointing to 10.227.1.122 & 10.227.1.118. Both your 4500 switches are doing layer 3. You use those 2 4500 (layer 3 gateway) switches to get to that host.

When you run a traceroute the reason why you are seeing the packets alternating btw those 2 4500 (layer 3) switches is because all traffic originated by the router is process switched. Process switching does per packet load balancing.

STP isn't relevant in this case as we are talking about layer 3 and traceroute doesn't have anything to do with it.

[jeff.vargas]

That makes perfect sense and answers my next question, Why was my CPU utilization so high on the router when backups ran???

Do you have a suggestion on a preferred switching method? I need to get away from the process switching.

Thank You very much for the answer!

[sundar.palaniappan]

Cisco recommends enabling CEF switching. Warning: There are some CEF related bugs on certain IOSs.

You might already be doing CEF or fast switching. Do a show ip int (int #) and it will tell you what switching type is in use.

Good Luck!!

--Sundar