routing default gateway

[gag_1]

Hello:
I have two rp3440 RP-A and RP-B,interconnected with router ,and geographically at different sites.
RP-A (10.0.198.242)...Router-A(10.0.198.241)...WAN...Router-B(10.9.3.193)...RP-B(10.9.3.194)
We are testing fail recovery "SW" between machines. One is active and the other standby.
We will focus on RP-A network configuration.
RP-A (lan0 = 10.0.198.242) is on subnet 10.0.198.242/255.255.255.240
RP-A (lan1 = 10.0.198.226) is on subnet 10.0.198.224/255.255.255.240

RP-A> route to access RP-B (10.9.3.194) is:
10.9.3.192/255.255.255.240 10.0.198.241 UG 0 lan0 0
Beside this, and based in customer requirements ,we have a default gateway on lan1 to access services.
RP-A> Default gateway is lan1 (10.0.198.226) :
default/0.0.0.0 10.0.198.225 UG 0 lan1 0

Now, the problem:
I launch a ping from RPA (10.0.198.242) to RP-B (10.9.3.194) ,ICMP packet are sent using lan0 (Router-A).OK
I do a cable unplug, and I "wireshark" that ICMP packets are redirected to lan1, because it is the default gateway & lan0 is KO.
I assume that this is the correct behaviour.
After 10 mn , I recover the situation on lan0, plug the cable back.lanadmin status for lan0 is OK.

But,the amazing point , is that pings still going using default router in lan1......
traceroute confirms this point:

RP-A,sys,root # traceroute RP-B (go to lan1)
traceroute to RP-B (10.9.3.194), 30 hops max, 40 byte packets
1 10.0.198.225 (10.0.198.225) 0.487 ms 0.310 ms 0.263 ms
2 10.0.198.225 (10.0.198.225) 0.359 ms !H * 0.389 ms !H
The situation is recovered if I reboot the machine....

[Shibin_2]

Do you have any alternate gateway defined ?
What's your HP-UX version ?

[Matti_Kurkela]

In HP-UX, if you have more than one network interface, you should always use traceroute with the '-i' option to explicitly select the network interface you wish to test. Otherwise the results can be misleading.

Please see this thread:

http://forums.itrc.hp.com/service/forums/questionanswer.do?threadId=1171990

MK

[gag_1]

Do you have any alternate gateway defined ?
What's your HP-UX version ?
Only one default gateway on lan1.Hpux is 11.11
We don't use multihoming.We have reproduced the problem in our lab. In our lab we have:

RPA,sys,root # netstat -nrv
Routing tables
Dest/Netmask Gateway Flags Refs Interface Pmtu
127.0.0.1/255.255.255.255 127.0.0.1 UH 0 lo0 4136
10.0.198.226/255.255.255.255 10.0.198.226 UH 0 lan1 4136
134.134.134.65/255.255.255.255 134.134.134.65 UH 0 lan0 4136
134.134.134.64/255.255.255.240 134.134.134.65 U 2 lan0 1500
10.0.198.224/255.255.255.240 10.0.198.226 U 2 lan1 1500
159.23.14.0/255.255.255.0 134.134.134.66 UG 0 lan0 0
127.0.0.0/255.0.0.0 127.0.0.1 U 0 lo0 0
default/0.0.0.0 10.0.198.225 UG 0 lan1 0

After lan0 switch off, ping are sent using lan1. When we recover/plug the lan0, ping's still going on lan1. Even, more, we get ping requests from the other side ,that are replied on the wrong lan1.

About traceroute with argument -i ,it makes feel confused in this case. I thought that the lan interface would be selected as a result of routing tables .
Gonzalo.