[CentOS] Read performance on software RAID1?
John R Pierce
pierce at hogranch.com
Thu Apr 12 16:55:14 UTC 2007
WipeOut wrote:
> Hi,
>
> Since the data is available in two places on a RAID1 array does the
> code in any way make use of that for improved performance by reading
> (different) data form both disks at the same time? Or is the one disk
> simply a slave that just receives copies of the data and is never
> really read from at all?
reads are distributed to both drives. the kernel raid implementation
is considerably more sophisticated about this than the typical
bios/driver based fake raid.
More information about the CentOS
mailing list