Linux nvraid 0+1 question

[Tonacatecuhtli]

My six drive nvraid 0+1 array only reads from the first three drives. It properly writes to all six and the array is healthy. Is this a deficiency of dmraid or the nvraid bios on my MSI P7N Dimaond motherboard?

Or, can anyone tell me if Windows properly uses all hard drives in a nvraid 0+1 configuration?

[harl]

I think it's normal.
Why should the Raid read twice the same info?
The mirror is only for redundancy

[Tonacatecuhtli]

I don't think that's quite it. Here's the scenario as I understand it. The data would look something like:

1 2 3 4 5 6
A1 A2 A3 A1 A2 A3
A4 A5 A6 A4 A5 A6

If we call the disks 1 to 6, from left to right (since my chart probably won't line-up,) the way it currently reads the data is.

1: A1, A4
2: A2, A5
3: A3, A6

It would be about twice as fast to do:

1: A1
2: A2
3: A3
4: A4
5: A5
6: A6

With this set up six blocks could be read simultaneously instead of three. Or is the thought that reading from the backup disks increases their chances of failure?

[Tonacatecuhtli]

Well, I hate to answer my own question, but the answer is that NVRAID can't read from RAID 1 drives simultaneously. The problem is that hard drives have a hard time reading one sector, skipping the next X and reading the next sector without having to have the platter rotate all the way around again.

With RAID 0 each hard drive can read consecutively and by the nature of striping will read every Nth (how ever many drives) sector, this is what we want. If this were attempted with RAID 1, either the same data would be read from each drive, or a large amount of rotational latency would be introduced as the drives read small bits and tried to seek to the next block needed.

Hopefully that'll help in case anyone else has the same question in the future.