Link to home
Start Free TrialLog in
Avatar of peter2407
peter2407

asked on

How to calculate storage by reducing RAID sizes?

My server environment and SAN is currently based mainly on RAID 6. I need to reduce this. How do I calculate what the reduction will be if I change, for example, a RAID 6 to a RAID 1 etc?
ASKER CERTIFIED SOLUTION
Avatar of Lee W, MVP
Lee W, MVP
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
So if you have a RAID 6 today with 6*100GB drives, then you have 400 GB of usable space.  If you go to a RAID 5, then you have 500 GB... RAID 1 is not possible (without wiping the drives and recreating the array - actually 3 separate arrays), RAID 0 would give you 600 GB, and RAID 10/0+1 would give you 300 GB.
SOLUTION
Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
SOLUTION
Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
gary, there's no difference.  math rules state multiplication is done before addition and subtraction.  So if we assume n=6 and the smallest drive = 100 GB, we can see that my formulas work just as well as yours.

n*smallest drive-2n
6*100-2*100 = 600-200 = 400, which is exactly the same as (6-2)*(100)
Yes, I'm well aware of the math rules [I did a PhD in Math :-) ] ... as I noted above, clearly you know what you meant, but the formula's aren't stated correctly.

To use your example:

n*smallest drive -2n  for n=6 and smallest drive = 100 ==>  6(100) - 2(6) = 600 - 12 = 588 :-)

What you SHOULD have said (to keep the order you used) was
n*smallest drive - 2*smallest drive   ... which is better expressed as simply (n-2)*smallest drive

Avatar of peter2407
peter2407

ASKER

Thanks all for the speedy response. The objective here is to reduce the volume of data being managed under a managed service for the SAN agreement, where we pay ber Tb. We have c.500 Tb and growing. There are 2 initiatives - reduce 'unnecessary resilience', therefore my incorrectly described label of 'reduce RAID'. There is also another initiative of de-duplicating data. Any recommendations there?
de-deplicating data.... then I'd suggest you look into Windows Storage Server - it is supposed to offer Single Instance Storage - that is, where if you have two copies of the same file, only one is actually the file - the other is merely a link to it... until and unless something makes it unique.

(Thanks Gary, that'll teach me to hop back on EE after my cat wakes me up in the middle of the night... just when I think I've got the math thing figured out.... well... I never was good at proofs (though oddly good at programming....).
StorNext from Quantum offers automated Hierarchical Storage Management and de-dupe. It's well groovy stuff: http://www.quantum.com/Products/Software/StorNext/Index.aspx. It's licensed by data under management. Archiving is also a good option. Symantec's Enterprise Vault will allow you to pull data off Tier 1 and bung it onto Tier 2 or 3. Again, well groovy stuff. http://www.symantec.com/business/products/overview.jsp?pcid=pcat_info_risk_comp&pvid=322_1

Out-of-control file system growth is something I've been dealing a lot with lately. The biggest impact that it has is that you simply can't get a full backup because there's so much data, so that leaves you with three options: de-duplicate the data (and there's only one on-array de-dupe technology at the moment: A-SIS from Network Appliance), move data to Tier 2 then Tier 3 storage and back it up less frequently, or pull data out of the environment using an archiving technology.
Thank-you