The Environment:
Windows 2008 Datacenter R2 SP 2 with Failover Clustering Role. 2 and 3 node clusters
I have inherited a couple Windows SQL Server Clusters and each experiences the same behavior.
Each cluster node has three NIC Teams (on HP hardware), One Team is the NIC for Production traffic, One is Heartbeat, and the third Team is the Management network. Each of these Teams is on a separate VLAN and is assigned a static IP with each VLAN using different IP ranges.
The SQL DBA also sets up a clustered Management IP for each SQL Instance on the cluster. These SQL Management IPs are on the same subnet as the Physical Management network. Of course each SQL Instance has an IP assigned for the SQL Instance also. So for the example server I am looking at now, there are 7 Management IPs with one being the real Physical NIC and all others being Setup within SQL Server.
The Problem:
Whenever Cluster resources are moved from one Cluster Node to another, we lose connectivity to all of the Management IPs.
All Cluster resources appear to come online, but we cannot access the Management Network addresses. If we disable and re-enable the Physical NIC Team, all connectivity is restored and is accessible until the resources are moved again.
We have experienced the same problem in most but not all of the Windows 2012 clusters that we've built.
One of my peers has opened a ticket with MS support and at this point nothing has been found to explain it.
Has anyone else seen this behavior?