Switches using wrong default gateway

[KodiakSystems]

Switches are configured with a gateway of .1 yet every trace starts at .2. This causes them to be unreachable through a VPN. What am I missing?

Boise_IDF_POESwitch(config)# sh ip route

                                                    IP Route Entries

  Destination        Gateway         VLAN Type      Sub-Type   Metric     Dist.

  ------------------ --------------- ---- --------- ---------- ---------- -----

  0.0.0.0/0          172.18.78.1     18   static               250        1

  10.61.23.101/32    172.18.78.1     18   static               1          1

  127.0.0.0/8        reject               static               0          0

  127.0.0.1/32       lo0                  connected            1          0

  172.18.78.0/24     VLAN18          18   connected            1          0

  172.25.78.48/28    eICU_WiFi       253  connected            1          0

Boise_IDF_POESwitch(config)# traceroute 10.61.23.101

traceroute to 10.61.23.101 ,

              1 hop min, 30 hops max, 5 sec. timeout, 3 probes

1 172.18.78.2           3 ms      3 ms      3 ms

2 Operation aborted.

[EricAtHP]

The 172.18.78.2 address actually comes in the response from the first hop router. Is it possible that the 172.18.78.1 is a VRRP address and the owner of that address is 172.18.78.2 at the moment?

[Vince-Whirlwind]

It might be easier if we could see the IP addressing config of the switch - VLAN interfaces.

if .1 is a firewall or router, it might not be setup to reply to ICMP.

[KodiakSystems]

The gateway for this switch is not using HSRP on the VLAN. This behavior only happens on about 5 switches out of 12.

Here is the config that is the same on all switches, with the exception of the IP address on VLAN 18.

ip default-gateway 172.18.78.1
interface 49
   speed-duplex auto-1000
   exit
vlan 1
   name "DEFAULT_VLAN"
   no untagged 1-32
   untagged 33-52
   no ip address
   exit
vlan 17
   name "DHCP LAN"
   untagged 1-32
   tagged 49
   no ip address
   exit
vlan 18
   name "static LAN"
   tagged 49
   ip address 172.18.78.12 255.255.255.0
   exit

[Vince-Whirlwind]

..and when you issue the ping command on that particular switch, the response comes back from .2?

What is connected to port49 on that switch? which port does it use and what is its IP configuration?

Where is .1 configured on the network?
Where is .2 configured on the network?

[KodiakSystems]

If I trace from the problem switches the first hop is .2 instead of .1.

Port 49 is the core switch. All 12 switches are connected to the core switch and only 5 have this issue.

.1 is on the primary router

.2 is on the secondary router

Both routers are connected to the core switch.

[Vince-Whirlwind]

What do you mean "secondary router"?

Write down your physical topology and trace the layer2 paths of your ping packets.

Then read this:

http://www.l3-switching.com/?p=397

[KodiakSystems]

The secondary router is a backup router with a backup internet connection. It will only be used when either the primary internet connection or router fails.

I'm not following why some switches would get an ICMP redirect and others on the same network with the same configuration connected to the same routers wouldn't.

[KodiakSystems]

I turned off redirects on the primary router and the problem still exists.