Re: btrfs data dup on single device?

[Konstantinos Skarlatos <k.skarlatos@gmail.com>]

On 25/6/2014 5:41 μμ, Christoph Anton Mitterer wrote:
> On Wed, 2014-06-25 at 08:47 +0100, Hugo Mills wrote:
>>     This has variously been possible and not over the last few years. I
>> think it's finally come down on the side of "not",
> I think that would really be a loss... :(
>
>
>>     The question is, why?
> Well imagine you have some computer which can only have one disk drive
> (laptop, etc.) and you still want at least some kind of redundancy
> against bit rot errors.
>
>
> IMO, btrfs should support most flavours out there...
> - n-way duplicates on the same device (and not just DUP with n=2)
For the same device there is also erasure coding, where you lose lets 
say 10% capacity, and have the benefit of recovering from the most 
probable disk errors that dont take the whole disk with them, bad sectors.

> - n-way mirrors on multiple devices (i.e. what we have right now with
> RAID1 plus up to classic RAID1 with copies on each device
> - RAID5/6
> - n-way striped+parity with n>2
> - "stacked" layouts (RAID 10 as e.g. MD has it,... RAID50, 60)
>
>
> And terminology should really be re-worked... IMHO it's very bad to use
> the term RAID1, if it's not what classic RAID1 does.
>
>
> Cheers,
> Chris.
>


-- 
Konstantinos Skarlatos