My company has had a machine set up for a while now under a domain name, and it has started to exhibit some truly baffling behavior. I won't use our actual domain, but I will use a placeholder for the purpose of examples... Let's say it's "exampledom.com". Anyway, here's the situation:

Over night, our nameserver disappeared. And when I say this, I don't mean half of the domains or half of the files, or stuff like that. In fact, all of the domains are in perfect working order and you can access them just fine. But the actually nameserver and main address (exampledom.com) is not working. A bad hosts file, you would assume, but no - that's working fine. So we looked into it further, and now it starts to get weird. You can actually access the www subdomain (www.exampledom.com) just fine, but all other subdomains (including the actual registered nameserver with GoDaddy, ns1.exampledom.com) don't work at all. All the hosted sites are working just fine though.

So to outline my problem more clearly, let's set up a little scene:
I have a server with four sites hosted on it: maindom.com, omg1dom.com, omg2dom.com, and omg3dom.com

Here's a list of what works and what doesn't:
*.omg1dom.com - WORKS
*.omg2dom.com - WORKS
*.omg3dom.com - WORKS
maindom.com - FAILS
www.maindom.com - WORKS
ns1.maindom.com - FAILS
*.maindom.com - FAILS

Understand what I'm trying to explain?

Anyway, we've checked all the records and zone entries, we've mucked about with GoDaddy, and what not, and all attempts have come up with nothing. This just happened out of the blue, so it really baffles me wtf is going on. Anyone have any ideas?

PS - Also, sorry if I'm a bit incoherent. I'm on hydrocodone at the moment, and am barely even awake.

If this is your name server and you have checked the zone file configuration, I would venture to guess this would be a network latency issue.

When this problem occurs is it normally during large spikes in traffic? MTR is a great way to quickly identify problem points within your network, http://en.wikipedia.org/wiki/Mtr_(My_traceroute).

This is the main reason why you should have at minimum 2 name servers which are independent of one another for failover purposes. How many do you guys have?

We just set up a second nameserver, and we were about to add it to the cluster via Webmin and as we were going to add it as a DNS slave and sync over the domain entries, we noticed that said addresses were not resolving.

And it's not a sporadic thing at all. It just started out of nowhere and we can't seem to fix it. We checked things out, and it resolves on the box itself, but when it does so, it uses the local loopback, so that kinda doesn't really help.