When physical disks are configured as the virtual disk, RAID group attributes assigned to the physical disks remain valid. In other words, the related attributes regarding virtual disk reliability and performance largely depend on the RAID group. Thus, it is necessary to plan how to use the virtual disk in advance.
Setting up a virtual storage pool for each RAID group makes it easier to recognize virtual disks when creating the disks for each purpose of use.
Guidelines for the RAID group for each purpose of virtual disk use are as follows:
Virtual disk for files or databases directly used by the transaction server
RAID1+0 (with value placed on reliability and performance)
Virtual disk for log files used indirectly
RAID1 (with value placed on reliability)
Virtual disk for backup or temporary use
RAID5 (with value placed on access load balancing)
When storage has an assigned a controller module (CM), you must determine the CM as the access path for each RAID group in order to balance the controller load. You can also consider the access path from the transaction server to determine the assigned CM. However, the copy function of the virtualization switch performs data copy processing between the connected ports of the switch and storage. Thus, if the CM in the storage connected to the virtualization switch does not assigned, the access performance of the transaction server via the assigned CM may be adversely affected.
