Raid 0 to Raid 5

[William Della Noce]

I have a ML370 with 4 hotswap 72GB disks confgured as Raid 0.
I have an aditional 72GB HDD and I want to configure a Raid 5 array without loose the data already stored on the 4 drives.
There is a way to do this?

Thanks in advance

[Greg Carlson]

William,
You have room for two more internal hard drives. Add one to two hard drives and then you will be able to migrate from Raid 0 to Raid 5. You can't change from raid 0 to raid 5 without adding at least 1 hdd because the available space on Raid 0 is greater than what Raid 5 will offer on the same platters. Once you add the hdd, you'll have the option to change to raid 5 in the ACU.

Ciao,
Greg

[William Della Noce]

Hi Greg

After pulging the new drive, and using ACU 6.40, It only can be used to expand an existing array or to create a new array.

I tried to migrate a Raid 0 array but it only allows me to modify the stripe size. The Raid option just show Raid 0.

[Greg Carlson]

hightlight the array you want to either expand or migrate (migrate in your case) then choose the hdd you want to add to the array,

Then when you create your logical drive you change the raid status. I believe you chose Extend Size option which only adds on the space and leaves your raid config the same.

[e4services]

William,

As stated above, you will need to add 2 drives to change a RAID level from any of the RAID0 to RAID5.
In review you have 4 RAID0 72GB drives all having data on them. Adding one drive will allow you to change the level of one of the RAID0's to RAID1, a mirror of 2, only.
It seems the confusion here is that you want to take 2 of the existing RAID0 drives and on new to make a RAID5 set. This would only be possible if you delete the 2 RAID0's first, loosing the data, then restoring the data to the RAID5 once you join the 3 together.
Of course that doesn't address the changes in the logical drives and how they were used (pointers to DISK3, DISK4 etc. or E:, F: etc.) from outside the data.

[Hein van den Heuvel]

The ACU Doc suggest it should be possible to migrate Raid-0 to Raid-5 through the "Migrate a Logical Drive" option.

From page 3-11:

"This option lets you alter the stripe size (data block size) or RAID level, or both, for a selected logical drive. There might need to be unused drive space available on the array for the migration to be possible, depending on the initial and final settings for the stripe size and RAID level."

Going from 4 drives Raid-0 to 5 Drive Raid-5 is theoratically possible, even if the logical drive fills the whole space. But I suppose there could be meta-data needs to through a monkey wrench into this.
Does your logical drive(s) fill all available space?
Unfortunately the manual text does not provide details on how much 'slop' is needed.

The manual does suggest
- make a backup
- it'll take a while.

You might be faster and safer to go for the roll-out, delete, recreate, roll-in method, but that may cause unacceptable down time?

fwiw,
Hein.

[Leslie Martin]

Which RAID Controller do you have ? Some do not
support online RAID migration

[William Della Noce]

I have a Smart Array 642.

Let me see I understand all that info:

1.- It's not possible to create a Raid 5 with the available hardware. I need to add 2 72GB hdd to the server to create a Raid 5.
2.- The best chance that I have right now it's to loose the data on to disks to migrate those drives to Raid-5

One question: If I reset all the disks configuration and start all over again, I will be able to create a Raid-5 array with 5 disk, using 4 drives to store data?

Thanks everybody for your help.

WDNV

[Hein van den Heuvel]

>>> One question: If I reset all the disks configuration and start all over again, I will be able to create a Raid-5 array with 5 disk, using 4 drives to store data?

Yes, ALL basic RAID-5 implemenations give up one disk to store redundancy = parity = xor data. That redundacy is not located on any single given disk, but spread over all, to balance load/avoid hotspots.

With 5 disks, you will have 4 disks worth of data available while being protected against single drive failure.

Hein.

[William Della Noce]

Thanks guys