Issue with DHCP migration

jamiehall5950 · 2014-02-15T19:43:45.000Z

Hi,

So we’re in the process of transferring / migrating from one server to another. So far we’ve managed to replicate roles and to add the machine as a second DC. I decided the next best option now that DNS and AD worked was to move the DHCP server from the older server to the newer one, transferred all settings (by hand), double checked and all seemed well. I then removed the scope from the old server and stopped the DHCP service.

I then executed a script remotely on a few clients to release DHCP and then to renew and that’s where it all has gone wrong. One machine out of the chosen few has gotten a DHCP lease but all the others have gone offline (guessing no DHCP lease given). The main problem is that I’ve done this remotely, am I going to have to go down to the offices tomorrow and renew / restart each machine or will windows automatically retry to gain a lease? The clients are running Windows 7 with 2/3 running Windows XP.

Thanks, Jamie

Gary-D-Williams · 2014-02-15T19:54:38.000Z

From reading your plan, it should have all worked.

Does the second DC have DHCP configured and authorised? Can you connect to the new DHCP server and see the logs? See if there are any errors, etc

jamiehall5950 · 2014-02-15T20:05:59.000Z

That is exactly what I thought, weird that one machine got a lease and the rest didn’t…

It is configured and authorized and I can connect fine. The logs show no issues and it does show that one machine getting a lease.

ryanrroberts7743 · 2014-02-15T23:47:33.000Z

Because one client came back, I will assume the method you used to script and execute was fine.

As I understand it, you did not restore the dhcp database from the old server to the new server, and you only created a new scope with the same IP range. One problem I have with that, is now you will be dealing with IP address conflicts. Maybe you got lucky with the one that came up, but the others may have pulled IP addresses that are already in use, causing IP address conflicts.

That would be my first guess.

jamiehall5950 · 2014-02-16T06:12:42.000Z

Hmm yeah that sounds about right, going to pop in and turn off the switch and turn it back on, hopefully forcing all the machines to renew…

mikedavis · 2014-02-16T11:33:43.000Z

How did you issue the renew command after you issued the release command? Did you copy a batch file down locally and call it? As long as you DHCP server is running fine, if the end user restarts their machine it should come back.

jamiehall5950 · 2014-02-16T12:02:25.000Z

Well I sorted it manually now but that could of been the reason! I didn’t think about running it locally. I ran it off the network and that probably released it, disconnected and then stopped reading the script

Gary-D-Williams · 2014-02-16T13:59:01.000Z

Ahh right, if you did an ipconfig /release from a script on the network it’ll never complete the script because it can’t read it from the network.

ipconfig /release = fall off the network

emester · 2014-02-16T22:13:42.000Z

restart the DHCP SERVICE, REBOOT A PC, if it fails to get the correct up address stop the DHCP SERVICE on server, reboot a pc - if it gets an address see what is handing out addresses. if you still have some getting address and some not then I would power cycle switches and routers between DHCP server and pc not getting an up address. hope this helps

michaelrosado2737 · 2014-02-17T03:10:10.000Z

I would have created a test subnet on both the old & new DHCP servers. Take advantage of lease time. Set the lease to one hour on your test subnet, deactivate the scope on the old serve and activate the scope on the new server. After an hour, the test device will request an IP address and the new server should respond.

If that works, use the lease time on the old servers. Make it something like 72 hours. Check your leases. deactivate one scope at a time while activating the new scope.

Mike R.