Hi,
I am changing a standalone SQL Server 2008 R2 to a Windows Failover Cluster with SQL Always On Availability.
This server has the following drive setup:
On the planned Cluster servers will have the same drives.
Should each drive have it's own vmware virtual data store?
Or should there be separate vmware data stores for the SQL Data (E:) DatstoreA and SQL log files drives DatastoreB and the rest (C:,
and G:) be on one data store DatastoreC? So three different datastores in total?
What is the recommended or best practice?
The links in this thread should help: SQL in Cluster
Hi, 4DAJ.
It really depends on your storage capabilities and IOPS requierements.
It's always better to manage less datastores as possible from a vSphere perspective. Also it's easier for migrations and configurations but not much really.
You may take in consideration that some disks requiere more performance than others and the less workloads you have on tier 1 datastores, the better they will perform as they won't be fighting for resources.
I would recommend having the RW intensive disks in SSD or SSD NVMe drives, while the other ones in mechanical or tier 2 datastores if you don't have enough space in the tier 1 datastore
Thanks, I decided to go for one datastore.
I will be configuring the unit allocation size of 64K for SQL data, logs and temp drives.
Would this cause any issues on the one datastore, the C: drive is set as default for the unit allocation size?
