Users Guide

4. Use disk shares to distinguish high-priority VMs from low-priority VMs.

NOTE: Disk shares are relevant only in a given host. The shares assigned to VMs on one host

have no effect on VMs on other hosts.

5. Run the applications to determine whether or not VM performance is acceptable.

Using the adaptive scheme to make virtual disk decisions

1. Create a large virtual disk, such as RAID 1+0 or RAID 5, and enable write caching.

2. Build a VMFS datastore on that virtual disk.

3. Place four or five virtual disks on the VMFS datastore.

4. Run the applications to determine whether disk performance is acceptable.

If performance is acceptable, you can place additional virtual disks on the VMFS datastore. If

performance is not acceptable, create a new, larger virtual disk, and repeat the process. You can use

a different RAID level. Use migration so that you do not lose VMs when you recreate the virtual disk.

VMware ESXi host fibre channel configuration

The MD vCenter Plug-in allows an ESXi host to be automatically configured to use a MD storage array by

detecting the installed host bus adapters (HBAs) in the ESXi host and configuring new hosts on the

storage array with the World Wide Names (WWNs) of the HBAs from the ESXi host. The default ESXi

multipathing mode for MD Fibre Channel storage arrays is Most Recently Used (MRU). To ensure

optimum performance for the ESXi host with more than two HBAs, you must configure ESXi host to use

the storage array in pairs of HBAs. This method allows for the maximum I/O throughput from the ESXi

host to the storage array. Using this method requires proper SAN configuration and balancing of LUNs

between hosts/host groups. Figure 70 shows a properly configured two-HBA port, ESXi host, SAN

configuration utilizing two fabric switches and a dual-controller storage array.

Figure 68. Dual port HBA configuration