I have been struggling to get a new server accessible on an existing network for about a week now. I keep running into a problem with my network's dns. Here's the situation:
A server named "proj" went down and needed to be rebuilt. I have done so using identical hardware but with a new os, CentOS 6.2. I finally have the new server in a acceptable working state. It is receiving a dhcp lease from "pdc" (Primary Domain Controller, which is also dns/dhcp server), I can ssh to and from, and have even had an rsync transfer going for about 24 hours now. In a Linux environment I am happy.
I've started to try to get it accessible over windows using samba. I've used other working servers as guides for smb.conf, here's the relevant excerpts:
Code:
[global]
workgroup = BULLSEYE
server string =
netbios name = proj
security = user
passdb backend = tdbsam
#============================ Share Definitions ==============================
[proj]
comment = project directory
path = /proj
browseable = yes
public = yes
guest ok = yes
force user = render
force group = users
read only = No
As far as I can tell there's nothing wrong with it, but I still can't see the machine from a windows box.
I'm pretty sure the dns server is setup to automatically update (at least I've seen it happen for other machines) but it hasn't added any entries for this new server. I'm guessing it is because it has the same host name (proj) as the old server, which still has an entry in the named zone files. Before touching anything if I do "host proj" it returns the old ip address. When I update the files it says "Host not found: 2(SERVFAIL)"
If this were the only issue it wouldn't be so bad, but once I update the named files things start to go bad. Not immediately, but eventually certain machines on the network become inaccessible through certain protocols. (ie can't connect to mail through snmp, but can still ssh to server; samba shares go offline; etc)
I haven't been able to spend much time testing things out when the network gets to this state because it's a panicked rush to get everything working again. The only way I've been able to fix the issue is with a backup restore of /var/named and /etc minus /etc/samba (which is pretty large).
Since testing for solutions makes things very unstable, I want to hear some experts advice on where I might want to start looking for the problem. Could dns be setup to only update automatically, and my tampering is breaking everything? How can I remove the record of the old server?
Thanks in advance, this is just something I haven't encountered before when adding machines to this network, and I did not setup "pdc" so I don't know if there is any tricky configs in there that I don't know about.
