Hello All,
This is a backup/cluster/SQL resource question. My expertise is in general server administration and backups. I understand enough about SQL to give the appearance I know what I'm talking about, so please excuse any incorrect terminology.
Currently, our agent based backup software calls to the SQL Cluster IP (or AGL IP, if enabled) to initiate the backup stream. This means that all backup traffic traverses the 1Gbps NIC on the current-active node. This works - but it's slow and
impacts prod. We acquired 10Gbps NICs and installed them to the physical cluster nodes. The plan was to leave Prod traffic on the 1Gbps NIC, and then force backup traffic over the 10Gbps NIC.
Due to limitations of the backups software, there is no way to force the backups to use those 10Gbps NICs - because they're not associated with the Cluster IP - which is how the backup software initiates the stream.
General setup of our SQL Environments:
Standard Cluster (no AGL) - 2 physical nodes; Win2012R2; SQL2014; Cluster IP and node IPs(1Gbps NIC) are on the same subnet
AGL Clusters - comprised of a standard cluster (cluster IP) and a SQL node at offsite DR location (10Gbps WAN) AGL IP resides in same subnet as Cluster IP and cluster node IPs.
*note - If these were stand-alone SQL servers, this wouldn't be an issue for the backup software. The clustering is what creates the issue.
After some internal discussion and discussions with the Backup Vendor, we're looking to simply take the IP from the 1Gbps NIC and assign it to the 10Gpbs Fiber NIC - Then disable the 1Gbps NIC. The result will be all prod and backup traffic using the
10Gbps connection. Of course, we'd do this in a controlled manner - stopping SQL services and applying to passive nodes, etc.
The questions/concerns: Is this a supported "upgrade?" What kind of trouble should I expect if I do this? Will SQL balk at this change?
Apologies if I've left this vague - I tried to keep it simple and to the point. Thank you for reading.