I am in the process of removing our sole Windows Server 2003 R2 server (Nemesis) which was the DC that held all the FSMO roles. I have not removed it as a global catalog server or demoted it yet. I have an existing Windows 2012 Server (Prometheus) that is a physical server that now hosts all (5) FSMO roles. I just introduced a second Windows 2012 Server (Chronos) as a VM, and everything seems to be replicating and running fine. DNS is active directory integrated as well.
I followed the steps outlined in the link below. I stopped at the bottom portion titled "Removing the 2003 Windows Server from the Global Catalog Server".
I decided to review how our Exchange 2010 Server was configured to communicate with our DCs by running command "Get-ADServerSettings |fl", and I notice most everything pointed to the Windows 2003 Server. So I used the following article as reference to use the Windows 2012 Server that now holds all the FSMO roles.
After running the command "Set-ADServerSettings -PreferredServer prometheus.domainname.loca
l all the parameters now point to this server with the exception of DefaultPreferredDomainCont
rollers still points to the Windows 2003 Server.
I am just wondering if that is normal. Maybe it will update to the new server holding all the roles after I run a dcpromo on the Windows 2003 Server. This parameter may not even be much of an issue. So before I remove the final Windows 2003 Server from the domain, can you please let me know if this is even an issue or not. I attached a screenshot so you can see what I am talking about.
I simply want to remove the Windows 2003 Server and not have its removal affect active directory stability or cause mail related DNS issues, etc. I am still reviewing the event logs on the all the DCs and the Exchange server to confirm AD is healthy and determine if there are any communications issues. Yesterday morning I promoted the second Windows 2012 Server as a replicating DC, and I also transferred all the FSMO roles as well. I plan on removing the Windows 2003 Server next Friday to give it plenty of time to replicate everything, and reboot everything to check all the event logs and overall health