topic Re: default geteway failed in Operating System - HP-UX

default geteway failed

Steiner — Wed, 30 Oct 2002 10:32:12 GMT

Hello,

I have a L2000 server with the OS HP-UX 11.0. Unfortunately the routing via default gateway with a virtual ip address is failing. Only physical IP addresses are working. If I enter a virtual IP as default gateway, a telnet session works only once. But if I try to open a session again the rout to the server is not available.
Please, can anybody help me?
Thank you very much.

Martin

Re: default geteway failed

Donald Kok — Wed, 30 Oct 2002 10:40:19 GMT

Hi Martin,

Eh, why do you not want to use a real gateway IPadres?
Maybe you can use the IP of the machine itself as a gatewayIP. This works on another OS, but have not tried it on HPUX.

Goodluck
Donald

Re: default geteway failed

Adam J Markiewicz — Wed, 30 Oct 2002 11:07:44 GMT

Is the network with this virtual IP acessible directly?
Check your routing with: netstat -rn

You should have something like:

127.0.0.1 127.0.0.1 UH 0 140508 lo0 4136
172.23.1.67 172.23.1.67 UH 0 2872817 lan0 4136
172.23.1.0 172.23.1.67 U 2 0 lan0 1500
127.0.0.0 127.0.0.1 U 0 0 lo0 4136
default 172.23.1.254 UG 0 0 lan0 1500

If the network is not accessible directly, you're unable to route to the gateway.

Good luck

Adam

Re: default geteway failed

Ron Kinner — Wed, 30 Oct 2002 18:13:11 GMT

Also with 11.0 dead gateway detection is on by default. If you can't ping the gateway HPUX 11.0 will think it is dead and refuse to use it. It will work for a few minutes after a reboot but then stop.

ip_ire_gw_probe_interval:

Controls the probe interval for Dead Gateway Detection.
IP periodically probes active and dead gateways.
ip_ire_gw_probe_interval controls the frequency of probing.
With retries, the maximum time to detect a dead gateway is
ip_ire_gw_probe_interval + 10000 milliseconds. Maximum time
to detect that a dead gateway has come back to life is
ip_ire_gw_probe_interval. [15000,- ]
Default: 180000 (3 minutes)

Set it to 0 to make it stop:

ndd /dev/ip ip_ire_gw_probe_interval 0

then if that fixes your problem add the setting to /etc/rc.config.d/nddconf

TRANSPORT_NAME[0]=ip
NDD_NAME[0]=ip_ire_gw_probe_interval
NDD_VALUE[0]=0

so that it will stay after a reboot.

Ron

Re: default geteway failed

rick jones — Thu, 31 Oct 2002 18:23:32 GMT

i'm not sure that one can set the ip_ire_gw_probe_interval to a value of zero, however, one can set ip_ire_gw_probe to a value of zero to disable dead gateway detection.

the other option is to get the router(s) supporting the virtual IP address to respond to the ICMP echo requests :)

Re: default geteway failed

Ron Kinner — Thu, 31 Oct 2002 19:13:32 GMT

Rick,

You are right, of course, about ip_ire_gw_probe being the correct parameter. I didn't remember exactly so logged on to my local hpux 11.0 and asked it:

ndd -h |gw_probe

and it only came up with the _interval parameter so I assumed that was the correct one. Should have realized that the range statement at the bottom said you couldn't go below 15000.

For some reason (guess we are missing a patch or three) the 11.0 that I have a login on does not offer the correct option.

Ron

Re: default geteway failed

rick jones — Thu, 31 Oct 2002 19:26:12 GMT

dont feel bade, the ip_ire_gw_probe does not appear in ndd -h output on my more or les current 11.0 system either :)

ndd -h does not completely lift the curtain on things in ndd. ndd /dev/ ? (as described in the ndd manpage so I'm not giving away a secret :) will show everything, but give no help