[clue-tech] RAID no. of disks

[chris fedde]

Big arrays from EMC and others typically allocate drives into raid groups.
Over say 48 drives in a drive shelf you might configure 6 raid groups of 8
drives each.  The array management software then allows chunks from all the
raid groups to be striped into logical units that are presented to the host.

Remember that raid 5 uses a three block rolling parity scheme.  The data on
any single drive can be re-constructed from the two adjacent drives in the
raid group.   So the number of drives in the group does not impact the
ability to reconstruct from a single drive failure.

If someone did create a raid 5 group with 1000 drives it it you would indeed
get 999 drives worth of capacity.   But other practical limits tend to get
in the way of that big of a raid group deployment.

On Wed, Apr 8, 2009 at 8:41 PM, Angelo Bertolli
<angelo.bertolli at gmail.com>wrote:

> I know that things like RAID 5 need at least 3 disks, and can lose 1 disk.
>  The biggest RAID sets I set up are RAID 6 with 12 disks, and honestly
> that's pushing it a bit for fault tolerance.  But it does seem to work.
>
> Does anyone on the list know if there is an upper limit to the number of
> disks?  I am sure there must be, as the parity can't possibly be rebuilt
> from just using one extra drive, if you have 1000 data drives, for example.
>  I'm sure there has to be a tradeoff somewhere, like if I have 1000 drives,
> I'm not getting a capacity of 999 drives, maybe more like 900 drives.
>
> Angelo
>
> _______________________________________________
> clue-tech mailing list
> clue-tech at cluedenver.org
> http://www.cluedenver.org/mailman/listinfo/clue-tech
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://cluedenver.org/pipermail/clue-tech/attachments/20090409/5bfa639e/attachment.html