Hi all,
I implemented at one of my Customer a Windows cluster 2012 for SQL Server AlwaysOn and we faced a problem during a failover test scenario but I just want anyone to confirm if the behavior is normal or not and why.
My configuration is the following :
2 cluster nodes with Windows Server 2012 and a FSW. We don't have a shared storage for this cluster.
We have two network cards on each server. One card ("public") is used for cluster and client communcation and the other ("private") is used for cluster network communication only. The communication between the 2 nodes and the FSW (located on a fileshare server on the same domain) are made by using the "public" card.
We tested different failover scenarios :
1- Shutdown one node --> the failover of the core resources (FSW, name and ip of the cluster) is ok
2- Turn off one node --> failover ok
3- Disable the "private" card on one node --> cluster is ok
4- Disable the "public" card on one node --> no failover occurs and all the core resources are in the failed state
I don't understand why in the situation number 4 the failover doesn't occur. If I issue a manual failover of the core resources the failover works well. If I disable the "private" card after disabling the "public" card on one node the failover also occurs. Is this behavior normal for you ?
Thanks in advance for your responses
++
We have
MCDBA | MCITP SQL Server 2005 / SQL Server 2008 | LPI Linux 1