Using Red Hat WS 4: I built two Redhat environments which are working appropriately on their network but which I can not ping from other machines (windows or other linux machines). By pinging I mean using their dns name - I can ping with an IP. I built a Mandriva machine which I can see from windows boxes and other linux boxes. Although to tell you the truth I used the Mandriva 10.1 configuration tool and it did it all for me.

So I've been trolling around various websites looking for some solutions and everything I've tried so far hasn't helped the situation. Could someone point me to the correct way to start evaluating this issue and where I can look for some more concise documentation on what I should be doing?

Thanks
- Kevin

Here's some things I've already done:
1.) corrected hostname
2.) corrected /etc/sysconfig/network
3.) corrected /etc/dhclient-eth0.conf
4.) corrected /etc/resolv.conf

Well ... which machine is running the DNS? If you're using
a DHCP capable router you've probably configured a dhcp-
client file that's not being used ... check
ps -A | egrep "(dhc|pump)"
to find out which client the DeadRat machines are using.

They're probably not telling the dhcp-server their intended
dhcp names but are getting "random" names assigned.


Cheers,
Tink

Technically what you are saying makes a whole lot of sense to me. And this is what I've been thinking but keep in mind that i'm a newbie and not sure how to get around in Linux...

I did this on one of my "deadrat" machines

ps -A | egrep "(dhc|pump)"

and found two process in this list

1990 ? 00:00:00 dhclient
2450 ? 00:00:00 dhclient

Is what you are saying, I should not be running these services? Or maybe the fact that there are two running is a problem?

- Kevin

oh sorry ... one more thing

the dns server is actually managed by my company @ hq. I have no access. I'm in a newly formed remote office doing some porting work for this new company. All other machines seem to be working in this environment so I know the networking between the two offices is working. There's also another dns server in this office which is getting retired once the porting work is complete - I've tried to get my "deadrat" machines working with either dns without success.

Thanks
- Kevin

No, I wasn't suggesting you shouldn't be running those. I find it
odd that you're running two instances, and I haven't seen a
dhclient-eth0.conf before (the normal default name would be
dhclient.conf, God knows why deadrat changes things ;}).

The question, I guess, will be whether or not the dhclient.conf
has a statement like
send host-name "deadrat-1";
to tell the server what it wants to be known by ...


Cheers,
Tink

yes - there's one line in dhclient-eth0.conf:

send host-name "agatonerh.bea.com"; # temporary RHL ifup addition


I'm not liking the word "temporary"... but this is the one items I did change. Maybe I should copy this file to dhclient.conf. I checked my mandriva and it doesn't look like the mechanism is the same because they don't have the dhclient.conf.

Thanks

Which client do they (Mandriva) use these days? My
personal preference is dhcpcd ... I seem to recall that
they were using pump back in the day...

[edit]
Btw, I wouldn't rename it, chances are that RH scripts
are looking for that.
[/edit]

Cheers,
Tink

not knowing to much about this it does look like Mandriva uses dhcpd. There's a dhcpd.conf file in /etc. Although that looks like it configures the dhcp client it doesn't look like the machine name is stored there. The /etc/sysconfig/network seems to be the only place my machine name is defined in Mandriva.

This is similar to my RH where the netork file also contains my machine name.