Just to ensure I understand - the problem occurs when you split the WNLB cluster across two versions of Hyper-V? It works OK on 2008R2 and on 2012R2, but not when split?
From: [email protected] [mailto:[email protected]] On Behalf Of Miller Bonnie L. Sent: Tuesday, January 20, 2015 5:10 PM To: '[email protected]' Subject: [NTSysADM] Anyone successfully using Microsoft NLB in guests on Hyper-V 2012 R2 host servers? Just started the process of migrating guest machines from our 2008 R2 Hyper-V cluster to the new 2012 R2 cluster the weekend before last. Doing some legwork on this migration as we are also migrating to some new storage and I don't (currently) have an interim 2012 server for importing the VM config-might have to try getting one of those. Not a big deal, we don't have snapshotting or too much fancy config so I'm just recreating VM config and the NIC along the way-have done this more times than I can count with both 2008 and then 2008 R2. All servers are migrating fine, except when I go to set up a server that is part of a Unicast NLB cluster. Am I missing something, or does it just not work the same? I have two sets (each in 2 node clusters) and neither will work on the new 2012 R2 hosts, but work on 2008 R2 hosts with what I think is similar/identical prior config. Example: -2-node NLB cluster WS12 R2 Guest OS -Unicast mode, servers on the same subnet, cluster address on the same subnet. -One vm (working) is still living on the old WS08 R2 Hyper-V hosts. The other (not working) is on 2012 R2 Hyper-V hosts. -The guest machine will work "normally" (can ping/connect/etc) until I try to add back to the NLB cluster, at which time network communications seem to quit working. The virtual nic inside of the guest LOOKS like everything is correctly set & bound for NLB. -New virtual NIC IS set correctly to enable mac address spoofing. I keep going back to this as the problem looks just like what happens if you fail to have this set correctly, but it is there. Have even recreated the NIC again in the guest to make sure. -Problem is not dependent on failover clustering-it happens outside of the failover cluster on either individual Hyper-V host server. -Error message after configuring NLB will either be host unreachable or NLB not bound. -Deleted & recreated NLB cluster on old (working) guest early this AM to make sure it really is from the migration. That one recreates fine, new one has same problems. -Had IPv6 set up-removed IPv6 addresses from guests & unchecked boxes for bindings, recreated cluster with no IPv6 in play (short of disabling in registry) and get the same errors only on the migrated server. Does anyone have this working? Trying to figure out if I need to open a premier case as there could be some new setting in 2012 R2 interfering, or something not right with the host OS config/switch (although I'm not really using any of the fancy new features for switching.) We have several NLB clusters that will need to be migrated. TIA, Bonnie
