On Wednesday morning, DNS stopped working on one of the domain controllers, instead of sitting and reinstalling it, we had planned to migrate the DC away from this server, and onto a 2008 DC we had installed in the summer. So i took the opportunity to make it into the PDC for the domain, and in turn demote the 2003R2 server that was failing.
As we had installed the 2008DC in the summer, the ADPrep ForestPrep etc... Were all completed then.
Our setup (after adding the 2008 DC in the summer) was as follows
2003 R2 Primary DC (80% DHCP Scope)
2003 R2 Backup DC (20% DHCP Scope) (Dying Server)
2008 DC (No DHCP)
The intention was to make the 2008 Server the new PDC, so whilst everything was going wrong anyway, we took the opportunity to move the 80% scope to the 2008 DC, and make the non-dying server a BDC with 20% of the addresses
I did it in the following order
Transfer the DHCP scopes from the dying DC to the 2008 DC
Switch the “exclusion” ranges so that the 2008 server had 80% of the addresses available and the existing PDC had 20%
Transferred the FOSM roles from the 2003 PDC to the new 2008 DC
Made sure “Global Catalogue” was on the old 2003 PDC before decommissioning the dying server with dcpromo
Changed the IP Helper addressed for each VLAN on the Core switch to reflect the 2008 DC and the 2003 DC that remained.
Since then the main admin network (plugged directly into the core switch) works fine, and the ICT Centre on one of the VLAN’s works fine.
I’ve tested 3 of the other VLAN’s / SubNet’s and I’m having no luck getting an IP address at all when plugging into the switch. Ports are still untagged to the relevant VLAN they belong too; all that’s changed is the DHCP server
Changed all the scope options to reflect the DNS are now 2008PDC and 2003BDC.
Are your scopes configured in a superscope? We've had problems with DHCP helpers and superscope configurations here before, taking the scopes out of the superscope allowed clients to get the correct address based on where their dhcp request originated.
Turned out the initial cause of the problem, and continuing existance problem was a IP Phone with a switch built in.
Someone had plugged the PoE into the phone, and instead of plugging there laptop into the 2nd port, they plugged it back into the wall and it grinded the network to a halt.
The phone was DHCP'ing and im guessing the switch went mental as it wasnt an avoidable "im plugged into myself" problem which was what caused the server to get strained by a large amount of requests and eventually fall over.
Still, was a good opertunity to migrate to a 2008 DC, and its all good now. Worst 3 days ever :P
Teacher: can i have a bigger HDD on my laptop?
me: I expect so, how big is the old one?
Me: yes I'm sure, did you use the USB HDD I gave you?
Teacher: yes i moved all my programs onto it