I have built a simple test scenario for DHCP failover feature of Server 2012 R2:
DC1 - Server 2012 R2, DC and DHCP role
SV2 - Server 2012 R2, DHCP role
CL1 - Server 2012 R2, no specific role, DHCP client
There is one DHCP scope, created on DC1 with a lease time of 5 minutes (testing purposes). All three servers share the same LAN and IP subnet (i.e. no need for a DHCP relay device). In fact, all are VMs in a single virtual network.
DHCP failover relationship for the scope has been set up between DC1 and SV2: 50/50 load balance mode, MCLT = 2 min, state switchover interval = 4 min (again: these short times are for testing only)
Booting up CL1 for the first time:
CL1 gets valid ip config from SV2 with a lease expiring after 2 minutes (MCLT, as expected).
CL1 renews its lease after ~1 minute and gets a lease expiring after 5 minutes. From now on, CL1 renews its lease every 2.5 minutes without any problem, each lease expiring after 5 minutes.
After ~30 minutes, I disable the NIC on SV2, so it's unavailable on the network.
On the next DHCP renewal process of CL1, a lease is provided by DC1 and expires after 2 minutes (MCLT again, as expected). After 4 min, DC1 changes status to "partner down". From now on, lease renewals of CL1 will result in a lease expiring after
5 minutes - as expected. Although CL1 is still obtaining valid lease renewals, an event 1003 is recorded for each and every request ("Your computer was not able to renew its address from the network (from the DHCP Server) for the Network Card with
network address 0x000C29A0D1BA. The following error occurred: 0x79. Your computer will continue to try and obtain an address on its own from the network address (DHCP) server.").
After ~30 minutes I re-enable the NIC on SV2. It takes both dhcp servers about 1-2 minutes to get their failover state back to "normal".
From now on, CL1 is still receiving its lease renewals from DC1 (shouldn't this fall back to DC1 after some time?). And what's puzzling me most: Every lease renewal request leads to an event 1003 in CL1's event log (same text as above: error 0x79),
although all renewals are in fact successful.
Rebooting SRV2 or issuing "ipconfig -release" and "ipconfig -renew" does not stop these events from being logged every ~5 minutes.
The dhcp log on DC1 shows the following entry for each of CL1's renewal request:
36,10/15/15,14:18:52,Packet dropped because of Client ID hash mismatch or standby server.,192.168.40.101,,000C29A0D1BA,,0,6,,,,,,,,,0 36,10/15/15,14:18:57,Packet dropped because of Client ID hash mismatch or standby server.,192.168.40.101,,000C29A0D1BA,,0,6,,,,,,,,,0 36,10/15/15,14:19:01,Packet dropped because of Client ID hash mismatch or standby server.,192.168.40.101,,000C29A0D1BA,,0,6,,,,,,,,,0 36,10/15/15,14:19:10,Packet dropped because of Client ID hash mismatch or standby server.,192.168.40.101,,000C29A0D1BA,,0,6,,,,,,,,,0 36,10/15/15,14:19:31,Packet dropped because of Client ID hash mismatch or standby server.,192.168.40.101,,000C29A0D1BA,,0,6,,,,,,,,,0 36,10/15/15,14:19:35,Packet dropped because of Client ID hash mismatch or standby server.,192.168.40.101,,000C29A0D1BA,,0,6,,,,,,,,,0 36,10/15/15,14:19:39,Packet dropped because of Client ID hash mismatch or standby server.,192.168.40.101,,000C29A0D1BA,,0,6,,,,,,,,,0 30,10/15/15,14:19:47,DNS Update Request,192.168.40.101,CL1.contoso.com,,,0,6,,,,,,,,,0 11,10/15/15,14:19:47,Renew,192.168.40.101,CL1.contoso.com,000C29A0D1BA,,3794425134,0,,,,0x4D53465420352E30,MSFT 5.0,,,,0 32,10/15/15,14:19:47,DNS Update Successful,192.168.40.101,CL1.contoso.com,,,0,6,,,,,,,,,0
On SV2, there are no entries at all for the given time period shown in DC1's log above!
So my two questions are:
Why is CL1's lease not transferred back to SV2 after it has come back online?
Why is CL1 still not able to get a renewal in a "normal" way (and therefore continuously logging event ID 1003)?
Thanks for your help!
Matt