Windows 2012 DHCP Failover "Lost contact with partner"

Question

_{Wednesday, June 12, 2013 6:22 PM}

We have stood up a pair of Windows 2012 DHCP servers and configured DHCP failover (in Load Balancing mode) as per the Microsoft Technet article.  The two servers are on the same subnet, and time is in sync between them.  Almost immediately after configuring scopes to failover, one node loses contact with the other, and the alternate node goes to failover mode "Recover".

We have restarted both servers and services.  Both DHCP servers show as available.  Clicking "Replicate Failover Scopes" completes successfully with no errors, and the only two warnings in the Event Log are about credentials for Dynamic DNS updates and a dynamic IPv6 address (which we don't use at all.)

How come they can replicate, but have "lost contact"; and how do we get the one out of "Recover" mode?

All replies (5)

_{Friday, June 21, 2013 12:15 PM}

You should see Recover state only if one of the DHCP server actually went down and came back up. In this case the server which comes back up would be in Recover state but will transition to Recover_Wait state and come back to Normal after it syncs it leases. This should be relatively quickly. Did your servers continue to remain in the Lost Contact...\Recover state for long.

Did you see any events in the DHCP Server admin channel Microsoft\Windows\DHCP Server\Admin.

You can also enable AuditLogs (Microsoft\Windows\DHCP Server\AuditLogs) which logs all the server to server messages. You can enable the log by right clicking on Microsoft\Windows\DHCP Server. View->Show Analytical and Debug Logs. This will show the Microsoft\Windows\DHCP Server\AuditLogs in event viewer. You can then right click and enable the log and see if there are any failures being logged there.

_{Friday, June 21, 2013 4:28 PM | 1 vote}

It remained in this state for days after restart (for patching).  The events in the log were 20255 (lost contact with failover partner), 20254 (established contact), 20259 (state changed to RECOVER), and 20251 (state changed from STARTUP to COMMUNICATION_INT) repeatedly, within the span of a minute.

I deleted the failover relationship and recreated it, and it has been stable since.

_{Thursday, April 17, 2014 9:23 AM}

I too had to delete and recreate which seems to have sorted it.

_{Wednesday, July 2, 2014 9:32 AM | 1 vote}

I had the same problem and when I went to recreate the failover relationship I noticed that the other DNS server had the wrong IP address displayed in the dialog. I therefore added it by IP address just to be sure.

The cause of this may have been due to me changing the IP address of the server sometime after I'd originally set up the DHCP failover.

_{Thursday, June 18, 2015 12:53 AM}

Same here. Setup replication, tested and it was working find.

IP of replication partner changed, tried to replicate got error.

Restarted DNS and DHCP services, still erroring with DCOM error unable to communicate with server.

Restarted primary DHCP server, replication working at that point.

HTH,

David T