Hi SQLBuddy,
Thanks for your input. Unfortunately, I don't think it has been of help. Please see my comments, below...
sqlbuddy123 (4/14/2014)
...To begin with, you should have done I\O stress testing before moving on to the new cluster just to make sure if it can handle the workload.
...
This is a large SAN environment that doesn't just service one Cluster and was already 'live' and running production workloads when I got involved. Capacity of the SAN and Storage, both storage-wise and performance-wise, were good, running at 1/2 capacity, as I mentioned in the OP, and all of the Instances had light workloads. An attempt to 'stress test' was done by another DBA, but it was evident that this was next to useless because the SAN Admins would not let him 'stress' the system - for obvious reasons...it was already in production. The opportunity to carry out solid testing had been missed at the beginning of deployment of the SAN environment and so all we had to go on was performance metrics we could gather and to ensure we captured and analysed baselines both before and after deployment of each addition to the environment to understand the effect.
sqlbuddy123 (4/14/2014)
...Also check the Storage Cache, HBA Queue Depth and if SAN Zoning has been enabled.
...
Info on these areas is out there and was done based on that info. It's not so easy to find info on the questions I put in the OP.
sqlbuddy123 (4/14/2014)
...If the SAN Storage LUNS gets shared with other servers with different I\O patterns then it will lead to poor performance.
...
As stated in my OP, LUNs were matched one-to-one with a Windows Volume. There was no sharing, and again this info is out there, so no problems here.
sqlbuddy123 (4/14/2014)
...You should have a dedicated SAN storage attached to your SQL Server to have good performance.
...
Nope...as I mentioned in the OP, things have moved forward from the days of 'dedicated' SAN storage and have moved to a design where the Storage layer is made up of many Disc LUNs pulled together as MetaLUNs, and presented as one large monolith Virtual Array from which the SAN Admins carve out storage-based LUNs to present to the Servers, such as our SQL Cluster.
sqlbuddy123 (4/14/2014)
...And the path to the storage won't be much different for the data travel.
...
This is the bit I am interested in, but hasn't helped me much in understanding the impact. Sorry. 🙁