Re: Advice sought on moving to AMD64
On 6/20/05, Rory Campbell-Lange <rory@campbell-lange.net> wrote:
> Hi Everyone
>
> On 20/06/05, Lennart Sorensen (lsorense@csclub.uwaterloo.ca) wrote:
> > On Mon, Jun 20, 2005 at 03:49:33PM -0400, Nathan Dragun wrote:
> > > Woops, sorry about that bit of mis-information.
> > >
> > > What kind of performance gains have they been showing over the IDE/SCSI
> > > interface SATA drives, any clue?
> >
> > Well most drives have never hit the speed limits of the interface they
> > use, so in terms of raw throughput you generally won't see a difference.
> ...
>
> We intend to use 2 SCSI disks in RAID1 for the system and the others in
> RAID10 for the DB.
>
> There is (obviously) a lot of debate about SATA vs SCSI on the
> Postgresql list. The general opinion is that 7200 rpm SATA disks just
> aren't fast/smart enough to cut it for serious database use, 10K Raptors
> being a possible exception. Since SCSI drives are designed "to do
> physical I/O scheduling, because the CPU can issue multiple commands
> before the drive has to report completion of the first one. IDE isn't
> designed to do that..." [Tom Lane], we're going for that.
IDE wasn't designed for that, but indications are that there are plans
for SATA to be able to support tagged command queueing.
It hasn't been supported with early SATA hardware, which is a pretty
thin veneer of SATA interfacing atop otherwise IDE hardware.
But command queuing should be starting to be supported...
--
http://www3.sympatico.ca/cbbrowne/linux.html
"The true measure of a man is how he treats someone who can do him
absolutely no good." -- Samuel Johnson, lexicographer (1709-1784)
Reply to: