RAID5 question

[m00ds]

can someone tell me how much useable disk space is available in a 3x36 RAID 5 disks?? that's a combined total of 108GB, does that mean that 108GB is available for use?

[Karl]

RAID5 Capacity = (Number of Drives - 1) * Size of smallest drive

So in your case

(3 - 1) * 36 = 72 Gb

RAID5 works with a minimum of 3 disks - but it is advisable to use at least 4 drives, with 3 drives if one drive fails the whole array will go down, with 4 drives, if one fails the array will still work, but with slightly degraded performance.

[m00ds]

ok thanks Karl. that really cleared up things...

[m00ds]

karl one more question. if the useable disk space is 72GB, how would it look on a linux partition? would it show 2 paritions each with 36GB space or just one partition with 72GB space?

[Karl]

It would show as one drive to the system, so however many partitions you created.

[m00ds]

thanks karl

[dandanfireman]

[i] with 3 drives if one drive fails the whole array will go down, with 4 drives, if one fails the array will still work, but with slightly degraded performance. [/B]

This is acually not true. If you have a 3 drive array, If a drive fails, the array is still operable.

Raid 5 works by writing data to the disk in stripes. Lets say you have a 8k file. 4k of the file is written to the first disk. 4k of the file is written to the second disk and 4k of parity (redundancy data) is written to the 3rd disk. The parity information for different files will be on different disks. The second 8k file would be 4k data on disk one, 4k parity on disk 2 and 4k data on disk 3. If any of the disks fail, the data can be recreated by doing a calculation with the remaining data stripe and the parity information.

Like the illustration below

D = data stripe
P = parity stripe

Disks 1 2 3
File 1 D D P
File 2 D P D
File 3 P D D
File 4 D D P

[Karl]

Originally posted by dandanfireman

This is acually not true. If you have a 3 drive array, If a drive fails, the array is still operable.

Actually from personal experience I can tell you that you are wrong, RAID5 needs a minimum of 3 drives to opperate.

RAID 5: RAID level 5 is the most popular configuration, providing striping as well as parity for error recovery. In RAID 5, the parity block is distributed among the drives of array, giving a more balanced access load across the drives. The parity information is used to recovery data if one drive fails, and is the reason this method is the most popular. The disadvantage is a relatively slow write cycle (2 reads and 2 writes are required for each block written). The array capacity is N-1, with a minimum of 3 drives required.

That's taken from raidweb.com - just to make sure that what i've been taught at A-Level and on my degree wasn't wrong.

[dandanfireman]

Unfortunately sometimes what is taught in school is not always what exists in real life. It is true that 3 drives are required for a RAID 5 array. However, the mis-understanding is that you can still read and write data if one of the drives fails.

Let me be specific. You can create a Raid 5 array from 3 hard drives. It will work. If one of the drives fails, you will still have access to your data. There is no fault tolerant data advantage to more than 3 disk drives. Regardless of the number of disk drives, if you lose more than one drive....the array is dead.

This is not something from a class that I used somewhere along the way to get a degree. I have done this. I actually had a customer with a 3 drive array run for more than 1 and 1/2 years on a 3 drive array with one of the drives failed.

[rsaylor]

Very true, with 3 drives one drive can go DDD and you are still OK. However, not sure if the rest of the world does this, but IBM offers RAID 5E. Now, you can have 4 drives in a RAID 5E and if one drive goes DDD then a compreshion starts. If another drive goes DDD then you are in critical mode. Now you can have 2 drives DDD without loosing data. Then take it to a more advanced level, combine 2 RAID 5 arrays as a spaned array. I have been working with IBM ServeRAID since 1997.