Server 2012 not auto-registering with 2003 DNS

Question

Hope somebody can help with this, it's driving me mad.
I have about 600 VMs, all Windows Server of various versions from 2000 (I know...) to 2012. My Active Directory and DNS runs on Server 2003. Two of the six DCs (the DNS servers) are physical.
The problem I have is that a handful of 2012 VMs seem to stop registering in DNS. Most 2012 VMs are fine, and all the non-2012 VMs are fine. So based on that I think it's more to do with something being wrong with these few VMs rather than the
DNS servers. It's always the same servers that have the problem. It causes grief because once the DNS scavenge happens the servers get removed and then nobody can contact them by name.
The "quick fix" is to run ipconfig /registerdns on the problem servers, whereupon they put themselves back into DNS again but don't update the registration. So eventually they just get scavenged out again.
I'm wondering if it's something to do with how the VMs were created. Some but not all 2012 VMs were created from a VMware template and customised via vCenter. I've checked some of the servers for duplicate objectSid and objectGUID in Active Directory but
so far they're all different. There's nothing particularly obvious in the event logs on these servers, aside from saying they're failing to register IPv6 PTR records (which is correct, should probably turn IPv6 off...) but I don't think that'd cause IP4v A
record registration to fail?
Any suggestions? Thanks very much in advance.

Answer 1

Hello, thanks for the reply.
Yes the "Register this connection's addresses in DNS" box is indeed ticked. I tried un-ticking and re-ticking it to see if that kicked it back into life on one server but it made no difference.
Some of the servers that had the problem now seem fine, and are updating the DNS registrations daily and have been for over a month. Others seem to update for a period of time, maybe a couple of days, maybe a few weeks, and then stop. Others only update
when I do the ipconfig /registerdns.
Power Management is enabled, though these are Server operating systems, so I'd like to hope they wouldn't do anything as daft as shutting down the NIC (especially when they're trying to talk over it). However this must be a default setting as it's enabled
on all the VMs I've checked, and the vast majority are working fine. I'm going to try disabling it on one of the problem servers but I'm not terribly hopeful it'll kick it back into normal operation - unless you can point me towards a KB article or something?
I'd enabled the DNS Client Operational event log on some of the problem servers but to the untrained eye it doesn't contain anything terribly useful (I've yet to find out what the various return code and option values mean).