Raid 5?

jester77

I'm in the process of sorting out an NAS server. I was looking at using 4 x 250gb drives and was wondering how to set them up. I understand Raid 0 and Raid 1 but I'm not tototally satisified with them. One has no fall back if a drive goes and the other is expensive having a complete mirror. I've read about Raid 5 but I can't fully get my head around it.

Can anyone explain this in simple terms? I know that if I have 1 TB of space then 700GB would be available as storage. But let's say I have over 600GB of data, which I almost do, then would my data be guaranteed to be ok if one of the drives dies? Would it also be ok when I approach the 700 max?

PogMoThoin

These are from other forums. Dont know much about it myself.

5 DEFINITION Disk striping with parity
COMMENTS Parity data is distributed across with parity all drives in the volume. Normal data and parity data are written to drives in the stripe set in a round-robin algorithm, similar to RAID 4.
RAID 5 is multithreaded for both reads and writes because both normal data and parity data are distributed round-robin. This is one reason why RAID 5 offers better overall performance in server applications than either RAID 3 or 4. Random I/O benefits more from RAID 5 than does sequential I/O, and writes take a performance hit because of the parity calculations. RAID 5 is ideal for database applications

RAID 5 = Parity and data is striped accross 3 or more drives, you have the speed increase of RAID 0, but the due to the parity bits the array can loose a drive without data loss.
The Benifits = Good perfomance, and good redundancy
The Problems = Cost - 3+ identical drives and a mid to high end RAID controller are required

jester77

PogMoThoin wrote:

These are from other forums. Dont know much about it myself.

Nice pic, hehe!

I sort of understand what it's saying but I just want to be sure that the 300GB that would be used up for parity would be enough to cover a dead HD, especially when I get close to the 700GB capacity?

Khannie

You usually use one drive for parity in a raid 5 setup, up to a maximum of (can't remember, but someone else here will...think it's 8).

So....in a 4 disk setup, you're devoting 1 disk to parity. Raid 5 only becomes cheaper in a multi-disk setup (because the controllers are expensive). I'd consider it the best raid setup for a company scenario. If one disk fails, you just plug it out, stick another one in, the parity rebuilds the missing disk, and you have 0 downtime.

jester77

I was looking at the Buffalo Terrastations, for €700 I can get a 1TB NAS server which supports Raid 5. Sounds reasonable to me plus it's a nice sized box as well and very quiet. Don't think I would have a lot of change left out of €700 if I did the build myself. It's just the Raid 5 thing that's making me hesitant. Don't want a drive to fail and discover that I've lost all data.
What I don't fully understand is how the parity works? ie 1 parity disc supporting 3 discs of data?

Anti

Hi there, was reading your post earlier on and rembered seeing this link that might help you. its from www.tomshardware.com and is about setting up a " Build a Cheap and Fast RAID 5 NAS "

Dunno if this helps you at all, but i thought it might be worth a mention.

Anti

would help if i gave you the link eh.......

http://tomshardware.co.uk/2006/08/01/cheap_fast_diy_raid_5_nas_uk/

Wertz

To expalin the parity on all discs thing, it's best to illustrate it.



The parity of each block is stored on another separate disc....so if one fails and get's swapped out, the system can rebuild all the blocks and the original parity block form the data on the rest of the array. I think if you loose two drives simultaneously, that's the whole lot out the window though...

jester77

Thanks for the link Anti, looks good!

@Wertz:
Take the following scenario: 4 x 250 GB discs - 250GB used as parity. This leaves 750GB for data. Lets say I've 700GB spread across the arrray and a drive fails, lets say the second drive in the picture.

Can you explain how the data is recovered? Does the Cp contain all the data for C1, C2 and C3 and then A, B & D are recoverd from their paritys on the other discs?

Sorry, just a little confused

SickBoy

The RAID controller is able to calculate from the remain 3 drives what is missing. If the parity data is missing it recalculates it. If data is missing then it uses the 2 remaining data entries and the parity data to calculate the missing data entry.
Imagine it like simple math;

1+1+1+x=4

In order to calculate what x is it adds up the 3 1s and subtracts the answer from 4 to get the parity bit. This is an oversimplified version of how it works but you get my point...

jester77

SickBoy wrote:

The RAID controller is able to calculate from the remain 3 drives what is missing. If the parity data is missing it recalculates it. If data is missing then it uses the 2 remaining data entries and the parity data to calculate the missing data entry.
Imagine it like simple math;

1+1+1+x=4

In order to calculate what x is it adds up the 3 1s and subtracts the answer from 4 to get the parity bit. This is an oversimplified version of how it works but you get my point...

Yeah, I can see what you are saying, sort of makes sense when you put it like that! Thanks

steve-hosting36

Just to chime in that I have a terastation 1Tb unit and it's great. The 750gb is really 687GB with raid5, but the unit itself is very reliable, and you can easily remove the sata disks from the chassis.

SouperComputer

You'll know when a HD fails in a RAID 5 as the array will slow to a snails pace as the controller works overtime to reconstruct the parity! RAID 6 is even better, as it can tolerate 2 failures.

Just be sure when ordering the drives to try and get them from different batches. Otherwise if you end up suffering from a batch issue.............