Re: RAID + PostgreSQL?

On Mon, 2006-06-26 at 12:10, Guy Fraser wrote:
> On Mon, 2006-26-06 at 11:08 -0500, Scott Marlowe wrote:
> > On Mon, 2006-06-26 at 08:59, Tony Caduto wro
> >
> > I have to say this has NOT been my experience. With a pair of U320
> > drives on an LSI-Megaraid with battery backed cache (256M or 512M, not
> > sure which it was, it's been a few years) our pg server was noticeable
> > faster with a mirror set than with a single IDE drive (with cache
> > disabled) on the same machine. And faster than a single SCSI drive with
> > no RAID controller as well.
> >
> >
>
> Confusion Abounds.
>
> Mirroring does not improve performance. Mirroring writes the same data
> to two sets of identically sized partitions/slices, and reads the data
> from both and verifies the integrity of the returned data.

I know of no RAID controllers that do that with mirroring during reads.
The linux kernel RAID module also does not do this. While all writes go
to all drives, reads are normally interleaved between drives. You can
even use > 2 drives in a mirror set for improved read performance of
things like a data warehouse, where the loads are in bulk during off
hours and then many many users read different sections of the same data
sets at the same time.

If you've an example of a hardware or software RAID implementation that
reads from both drives and compares them to be sure the data has
integrity, I'd be interested in seeing them, as there are some uses for
that kind of setup, although you'd really need three disks, at least, to
assure a good read, since a failure in a two drive mirror would result
in a state where you knew the data was corrupted, but without check
digits no way of knowing which drive is failing.