TWO default routes configured by default

[Angel_11]

Hello hpux'ers,

I've a problem with two default routes configured on a HP-UX 11.00.01 system.
The system receive one request from the first GW, but it answer for the second route. The problem is that the Firewall discard this kind of messages for security matters.

How could I keep these two default routes and to force that one request receive from one GW will be answer from the same GW?

Thanks in advance.

ANGEL

[Michael Tully]

Hi,

I have a similar problem that I'm currently working on, (not fixed) but may solve yours.

# ndd -set /dev/ip ip_forwarding 0
if that doesn't work try
# ndd -set /dev/ip ip_strong_es_model 1

Have a look at the 'ndd' man page for further information.

Mick

[Santosh Nair_1]

This doesn't make sense...by definition you can only have one default route. Remember the default route is the catchall route. So if you define a static route to your internal network to go through your internal router and point the default route to go through the firewall, then anything destined for the internaly will go through your internal router and everything else will go through the firewall.

So if your network is
192.168.0.0 and your internal router is 192.168.0.1 and your firewall is 192.169.0.1 then add the route:

route add net 192.168.0.0 192.168.0.1 netmask 255.255.255.0 1

route add default 192.169.0.1 1

Hope this helps.

-Santosh

[U.SivaKumar_2]

Hi,
Only one default route will be active at a time. And the second route will be idle until
the system recognises that the gateway for first default route is not reachable ( dead gateway detection in networking terms ) and switchover traffic to second default route.

regards,
U.SivaKumar

[sven verhaegen]

Hi

I'm a bit surprised by this behaviour , normally the routing stack of the HP-UX doesn't allow usage of double default gateways before 11.x if multiple default gatewayss existed only the first was used even if the other side didn't answer any more , in 11.x however there exists a new feature called Dead-gateway-detect , if you have a gateway it will look for the activity on that gateway , at a 3-5 minute interval it will ping the gateway IP address and if it receives no reply (as with a firewall that doesn't answer to ping) it deactivates the gateway and if a second one exists it will use that one , before dead-gateway-detect , only the first default gateway was detected and the other just ignored

luckely one can deactivate dead-gateway detect by using ndd

ndd -set /dev/ip ip_ire_gw_probe 0

this de-activates the setting and only the first default gateway is used (as before), other gateway are also subject to this behaviour , only direct routes and host routes are not

if you alter this and it solves the problem , don't forget to make this pernmanent by adding hte parameters in de /etc/rc.config.d/nndconf file