topic Re: Network Access problem in Operating System - HP-UX

Network Access problem

Praveen Sharma_3 — Wed, 27 Nov 2002 14:10:55 GMT

Hi All,

I am facing problem on HP-UX 11.0 workstation (700 series machine) when i am trying to ping different subnet machine which is routed through Brick server(default gateway).

For e,g i am trying to access 198.180.37.48 machine from 164.164.27.210 machine which has a gateway(164.164.27.1)-Brick Server. Although It works for 1 minute then it is getting hang. When again i modified the gateway (164.164.27.1) for this machine then again it is started for next 30/60 secs.

Please somebody have some idea regarding this problem then suggest me.

Thanks,
Praveen

Re: Network Access problem

Jim Turner — Wed, 27 Nov 2002 14:15:48 GMT

Hi Praveen,

What does the output of a "netstat -rn" look like? Is your default gateway gone when the pings start to fail?

Cheers,
Jim

Re: Network Access problem

John Palmer — Wed, 27 Nov 2002 14:16:28 GMT

This could bu due to Dead Gateway Detection. Have a look at:
http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0xfaeecdec06f1d61190050090279cd0f9,00.html

Regards,
John

Re: Network Access problem

Praveen Sharma_3 — Wed, 27 Nov 2002 14:30:38 GMT

Jim,

No, output of a "netstat -rn" does not change.

Thanks,
Praveen

Re: Network Access problem

Adam J Markiewicz — Wed, 27 Nov 2002 14:44:33 GMT

I would check with 'traceroute' what happens before and after.
Mayby it's some problem in gateway?

Good luck

Adam

Re: Network Access problem

Praveen Sharma_3 — Wed, 27 Nov 2002 14:59:47 GMT

John,

I have executed "ndd -set /dev/ip ip_ire_gw_probe 0" command to disable dead gateway detection and add the entries into /etc/rc.config.d/nddconf file to permanently disable dead gateway detection .

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

But problem still remains same.

Thanks,
Praveen

Re: Network Access problem

Ron Kinner — Wed, 27 Nov 2002 15:31:12 GMT

OK. Lets check layer two first. Run
linkloop MACofRouter

And verify that it talks. It could be that your switch is going into errdisconnect mode because of a duplex mismatch or perhaps you have a duplicate IP address. Try a linkloop to the MAC of another host on the same network. Can another host talk to the router when you can't? If not check the switch port going to the router. Perhaps it has the duplex mismatch.

It wouldn't hurt to run
lanadmin
lan
display
(make sure you are looking at the correct LAN if you have more than one) and look for lots of errors especially on the second page. Could be you need to set your duplex and speed on both sides of the connection.

Are we running gated?

What does traceroute show before and after the failure?

Ron

Re: Network Access problem

John Palmer — Wed, 27 Nov 2002 15:37:10 GMT

Having disabled it, have you removed and recreated the route?

To confirm that dead gateway may be the problem, are you able to ping the Brick server?

Regards,
John

You only make mistakes if you're doing real work and getting things done.
Anonymous

Re: Network Access problem

Ron Kinner — Wed, 27 Nov 2002 19:00:27 GMT

I think we still may have the dead gateway detection problem. Seems like there is a patch that you need in order to make ndd turn off dead gateway detection. A previous poster ran into the same problem.

Ron