I had the similar result with IDE drives through hpt374 controller. Still
don't understand why a degraded system has better read performance.
YQ
----- Original Message -----
From: "Jakob Oestergaard" <jakob@unthought.net>
To: <raid@ddx.a2000.nu>
Cc: <linux-raid@vger.kernel.org>
Sent: Tuesday, October 22, 2002 3:24 AM
Subject: Re: Is Read speed faster when 1 disk is failed on raid5 ?
> On Tue, Oct 22, 2002 at 12:49:41PM +0200, raid@ddx.a2000.nu wrote:
> >
> > On Tue, 22 Oct 2002, Jakob Oestergaard wrote:
> >
> > > I didn't follow the beginning of this thread, sorry. Is this SCSI?
> >
> > Ide disks on 3ware 7850 using software raid5
>
> Ok. Odd.
>
> My first guess would be that it's some master/slave IDE issue - but you
> are probably running all your disks as masters (one disk per channel),
> right?
>
> Other than that... Well, if the parity information is intact, the disks
> would need to skip the parity blocks when the array is read from
> sequentially. With a degraded array, the reads are all contiguous on
> the disks - this could be a difference perhaps??
>
> What chunk size are you using? And can you try a chunk size that is an
> order of magnitude bigger or smaller? (might take some time to test
> this out).
>
> For example, if you use a 4k chunksize (I don't think you do, if my last
> guess holds), then try 128k. If you are using 64k or above (which,
> again, if I am guessing correctly, is probably more like what you're
> using), then try 4k.
>
> --
> ................................................................
> : jakob@unthought.net : And I see the elder races, :
> :.........................: putrid forms of man :
> : Jakob Østergaard : See him rise and claim the earth, :
> : OZ9ABN : his downfall is at hand. :
> :.........................:............{Konkhra}...............:
> -
> To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
>
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
