NIC confusion
 

Hi all,

I'm an experienced solaris admin who's totally confued by a Red Hat network issue.

Basically I installed (my first ever) RH Enterprise server on a HP DL360 with hardware raid....I had to create a backup machine exactly the same as the first...to save time I took one of the disks from the first machine (machine1), put it in the second (machine2)and rebooted....I changed the hostname and IP and all looked good.

Now for the strange part...if I run a tcpdump on both machines and ping machine1 from a third machine.....I can see the response comes from machine2....if I check the arp tables on the third machine, then I can see both IP's(of machine1 & machine2) with the MAC address of machine2, if I force a connection from machine1 to the third machine, then while the connection is kept alive, all is well (arp tables show correct info).....I'm very confused.

I rang the RH support line and was told "it's not a standard install...bugger off".

Any help would be MUCH appreciated.

Maybe the NIC configure file isn't correct!Check the NIC configure file in the /etc/sysconfig/networking-scripts/ifcfg-ethx(x:your nic number)

hi danielwupan,

I have chcked the /etc/sysconfig/network-scripts/ifcfg-eth0, it gives the standard IP stuff, but doesn't mention the MAC address. (an ifcofig - on machine1 and 2 - does show the correct addresses).

Anything else you can think of?

Boru

Can you please post:
/etc/hosts
/etc/host.config
/etc/resolv.conf

At first look, it indeed seems to be an ARP related issue. The IP addresses of both machines get wrongly translated into the MAC address of the 2nd machine. So your third machine sees the 2nd machine under 2 different hostnames.
However, if you consider that DNS can accomplish something similar (2 hostnames for 1 IP address), then you know that the problem may simply lie in the hostname configuration.
For instance, let's assume that machine 2 has been configured with 2 hostnames (which is unusual, but not impossible, I think). If machine 3 then broadcasts an ARP request to find the MAC address and machine 2 happens to respond first, it will give out it's own MAC address for both IP addresses.
Given that you've simply put the disk of machine 1 in 2 and added a hostname, you may have accidentally caused this. In any case, since you're not using DNS, your /etc/hosts files are used for hostname => IP translations.
So they should at least give a clue about what's going wrong.

Also, you could try doing a "ping" on hostnames and IP addresses of both machines to see if they behave differently.

You could also try looking in /var/log/* for error messages.

Hi timmeke,

here's the files requested:

/etc/hosts:
# Do not remove the following line, or various programs
# that require network functionality will fail.
###127.0.0.1 machine2 localhost.localdomain localhost
127.0.0.1 localhost.localdomain localhost
172.22.131.161 machine2

/etc/host.conf:
order hosts,bind

/etc/resolv.conf:
search mydomain.com
nameserver 172.22.150.161
nameserver 172.22.150.21


One other strange issue.....I had another admin look at the machine for me (he's in another part of the company)....he changed the broadcast address (to a completely wrong address) and disabled the broadcast function.....this "fixed" the issue (as as far as he tested it did).

I am convinced that I did cause this myself by swapping the disks between machines....short cuts!

Cheers,
Boru