On 04/29/2011 06:42 AM, Scott Marlowe wrote:
> I think you misunderstood. He's not storing 480GB on the drives,
> that's how much WAL is moving across it. It could easily be a single
> 80GB SSD drive or something like that.
>
Right; that's why you don't necessarily get saved by the fact that
larger databases must go onto more flash cells, too. Sometimes, yes,
but not always. The WAL is really close to a worst-case for flash:
lots of redundant information that's constantly overwritten. It's the
last thing you want to consider putting onto SSD. There's a good reason
why so many of the "enterprise" SSDs try to distinguish themselves with
redundancy and wear leveling advancements; it's so this sort of workload
doesn't kill them.
Combine that workload possibility with the limitations of MLC flash, and
you can see why the lifetimes actually are a serious concern in some
situations. Not all of them, of course, but this is why I recommend
things like directly measuring your WAL volume.
--
Greg Smith 2ndQuadrant US greg@2ndQuadrant.com Baltimore, MD
PostgreSQL Training, Services, and 24x7 Support www.2ndQuadrant.us
"PostgreSQL 9.0 High Performance": http://www.2ndQuadrant.com/books
