Routing anomaly in Catalyst 3750 network

Answered Question
Nov 8th, 2007
User Badges:

I have a problem I can't quite figure out.

If you look at the attached diagram:

If I run a traceroute from MRUCOLOPRODSW1 to a server on VLAN 20 off of switch MRUCOLOCORPSW1 It routes the packet through MRUMADCORPSW1 on VLAN 1 before returning to MRUCOLOCORPSW1.


Here's the traceroute:

MRUCOLOPRODSW1#traceroute 10.30.20.41


Type escape sequence to abort.

Tracing the route to 10.30.20.41


1 10.30.24.5 0 msec

10.30.19.101 0 msec

10.30.19.1 0 msec

2 *

10.30.20.41 0 msec 0 msec


Am I running too many instances of RIP?


Any guidance you could give me would be much appreciated.


Thanks,


-Rob Locke


I just ran traceroute again and got:

MRUCOLOPRODSW1>traceroute 10.30.20.41


Type escape sequence to abort.

Tracing the route to 10.30.20.41


1 10.30.19.1 0 msec 0 msec

10.30.24.5 0 msec

2 10.30.20.41 0 msec 0 msec 0 msec


MRUCOLOPRODSW1>


Correct Answer by Edison Ortiz about 9 years 8 months ago

In order to get to the destination host, it needs to go via the switch hosting the destination host, right ?


If you look at the mac-address-table on the sourcing switch, which Vlan has the MAC address of the destination switch first ? The table is read in order, first match wins.



  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (1 ratings)
Loading.
cfolkerts Thu, 11/08/2007 - 13:26
User Badges:

I would run RIP version 2 without auto summarization and repost the results.


router rip

version 2

no auto


Regards

robertlocke Thu, 11/08/2007 - 13:42
User Badges:

I ran RIP version 2 as per your instructions and still get routed through the other switch:

MRUCOLOPRODSW1>traceroute 10.30.20.41


Type escape sequence to abort.

Tracing the route to 10.30.20.41


1 10.30.19.101 8 msec 0 msec

10.30.19.1 0 msec

2 10.30.20.41 0 msec 0 msec 0 msec

MRUCOLOPRODSW1>



cfolkerts Thu, 11/08/2007 - 13:51
User Badges:

Per you drawing, this behaviour is OK since you have layer 2 links between the MRUCOLOPRODSW1 switch and MRUMADCORPSW1 switch. Try sourcing your trace from your vlan 24 interface on switch MRUCOLOPRODSW1. This trace should go through MRUCOLOCORPSW1.



Edison Ortiz Thu, 11/08/2007 - 14:02
User Badges:
  • Super Bronze, 10000 points or more
  • Hall of Fame,

    Founding Member

What you are seeing is normal. It's simply a behavior of the ARP and CAM entries in the participating switches.


Let's do a packet walk:


1) Packet is sent to the default gateway on that segment 10.30.24.5

2) The packet is Layer3 switched over to Vlan1. You have a trunk between the 2 switches and it will list all remote MAC addresses in order. I'm pretty sure Vlan1 will be seen first. (show mac-address-table)

3) On running the traceroute again, 10.30.19.101 won't respond because your gateway knows the egress port for the destination device. Do a show arp and it will be listed in your switch. If you wait a few minutes for the ARP to timeout, you will see the same first traceroute again.


HTH,

robertlocke Thu, 11/08/2007 - 14:18
User Badges:

OK, I get that. But why is it Layer 3 switched to Vlan 1 instead of Vlan 20, which is the destination network?

Correct Answer
Edison Ortiz Thu, 11/08/2007 - 14:25
User Badges:
  • Super Bronze, 10000 points or more
  • Hall of Fame,

    Founding Member

In order to get to the destination host, it needs to go via the switch hosting the destination host, right ?


If you look at the mac-address-table on the sourcing switch, which Vlan has the MAC address of the destination switch first ? The table is read in order, first match wins.



Actions

This Discussion