[1467] in linux-scsi channel archive
Re: Disk array latency theory (was: EATA performance issues)
daemon@ATHENA.MIT.EDU (Ingo Molnar)
Tue Feb 25 06:37:04 1997
Date: Tue, 25 Feb 1997 12:35:15 +0100 (MET)
From: Ingo Molnar <mingo@pc5829.hil.siemens.at>
To: Dario_Ballabio@milano.europe.dg.com
cc: linux-raid@vger.rutgers.edu, linux-scsi@vger.rutgers.edu,
lma@varesearch.com
In-Reply-To: <199702250827.JAA19306@milano.europe.dg.com>
On Tue, 25 Feb 1997 Dario_Ballabio@milano.europe.dg.com wrote:
> Disk syncing is a feature supported by some high quality disk arrays
> like the CLARiiON (see http://www.clariion.com).
what i'd >really< love to see is a 'head position reporting' feature. With
synced disks this is already kindof possible, if one syncs the host's
timings every now and then.
access >doesnt< have to be random. We are in the process of adding more
explicit IO scheduling to the current 'elevator scheduler' (as part of the
RAID project), and to have maximum optimization we need to know the head
positions. [the disk positions but it doesnt matter ;)].
It might even be OK to see the drift between disks, by some special 'query
head status' SCSI command or something like that. We can do highprecision
timings on most if not all Linux platforms, so this querying doesnt have
to be done only every few seconds or so.
inter-disk syncing is one good step, but unfortunately we do not know the
offset ... or is there some trick already available? ;)
we could actually calculate this offset by measuring data arrival
latencies, but this method has limited accuracy IMHO.
-- mingo
