HP Storageworks MSA 2212fc best practices configuration assistance

[Dave McLeod]

Dear All,

I have an HP Storageworks MSA 2212fc with 12 450GB disks, with 2 fabrics and 3 ESX hosts connected to it.

- Vdisk1 single volume 1.8TB
5 drives is configured with RAID 5 and

-Vdisk 4 volumes 2.2TB
6 drives configured for RAID 5 as well


This volumes are created for VMware datastores and production servers are currently running on them. We are also planning to P2V some more servers to the SAN which will require expansion in the near future.

We are considering implementing the following options

Options I
•    Expand the enclosure by installing a new MSA2000 Dual I/O 3.5" 12 Drive Enclosure
•    Create new vdisk group according to HP best practices
•    Migrate some VM to storage area network (SAN) disk created in Phase I above.
•    Recreation of the existing virtual disk.
•    Convert more physical servers to virtual and add to the new vdisk group.


Option II
Another option is to leave the current RAID group configuration as it is and implement the following:

•    Installation of a new MSA2000 Dual I/O 3.5" 12 Drive Enclosure
•    Expand virtual disk capacity by adding drives to it from the new enclosure.
•    Expand the volumes. The volume expansion does not require I/O to be quiesced; the volume can continue to be used during the expansion process.


Questions
Please could you assist with the advantages and disadvantages of both options, recommendations and if there are any workarounds



We also intend to P2V a file server that has 1.2TB disk will grow beyond the VMware 2TB limit (the file server is the reason why the 1.8TB was created)

Do we consider using extents if so what are the caveats (advantages, disadvantages, work arounds etc.)

OR

Create multiple 2TB hard disks inside the VM and convert then into dynamic disks and create a single volume inside Windows 2003 server

Please any manual, links that explains all these options will be very helpful.

[SamUK]

VMware do not recommend Dynamic disks to be used in windows for the VM's.

We had this challange for a customer who had a physical server with a single 4TB LUN - we decided to brake this server into smaller servers - this also helped with the businesses backup recovery times.

Also we tend not to have more then maybe 10VM's per datastore, for performance reasons.

so you could P2v, and Svmotion the servers to the new datastores when they before avalaible..