Strange behavior of interface lan0

[Alex_133]

We have a 9000/778/B180L(11.00) workstation where we recently replace its faulty mainbord, all the configuration remains same as we use back the original harddisk without reinstalling the OS.

Then we experince strange behavior of the lan0, for example: No other machine is able to ping this machine, but this machine is able to ping everyone.

When this machine establish a connection with some other machine(e.g. Telent/FTP session), then everyone is able to ping it.

I suspect when no TCP/UDP session is establish with other nodes, the OS will bring down the lan0, this may not be the case at all...

Anyone has experience this before?

[Mihails Nikitins]

Hi,

Looks strange...

You can check lan0 status at any moment by command
ifconfig lan0
I doubt it goes up and down...

Can you ping the host from itself (both loopback and public addresses)? Do you use DHCP for this workstation?

Do you ping the workstation from other hosts by IP address or by name? Try IP addresses first to be sure there are no problems with naming services.

BR,
Mihails

[Alex_133]

Sorry, this is a duplicate post.

[Alex_133]

....and shouldn't be under [databases] catagory, sorry again.

[curt larson_1]

in your network configuration files are you still using the MAC address of the old lan interface?

[G. Vrijhoeven]

Hi,

check out the speed and duplex settings on the card and on the switch. You can use lanscan to check the ppa number and lanadmin -x ppa ( ux 11) to display the settings on unix.

use lanadmin -X 10fd/10hd 100fd/100hd ppa to set the interface in the correct settings (100fd is prefered if possible.)

If these settings are correct check the netmask. (ifconfig lan0) settings


Gideon

[Paula J Frazer-Campbell]

Alex

If fine before the motherboard change then suspect the seating of the lan card.
Down the machine and remove and refit the card several times to clean the contacts.


Paula

[Ikuo Fukuda]

Hi alex.
Do you have same IP address host in your local lan?

[Alex_133]

Unfortunately, setting the lan0 to FULL DUPLEX didn't solve the problem. Also, the MAC address is not configured in the /etc/rc.conifg.d/hpbase100conf or /etc/rc.conifg.d/hpetherconf

We can isolate the problem from duplicate IP, DNS, cable or routers, because as long as I establish a TCP(e.g. telnet) to another node in the LAN, the others start getting PING response from this machine, whenever I terminate the session, then the others start having PING timeout.

[rick jones]

Actually, you may still have some duplicate IP or duplicate MAC issues.

When a TCP connection is established, the system will ARP for the remote IP address to MAC mapping. This is a broadcast and all the nodes in the broadcast domain will see it and update their IP-MAC mappings for the node. If they had a different MAC before, they would have been sending their pings to some other system which may have been silently discarding them or something.

So, as a triple check, examing the arp caches on the remote systems while you go through all those steps you mentioned, and see if the MAC addresses are as they should be.