Eliot Gable wrote:
> However, I have not seen much in the way of a serious attempt at
> multi-master write scaling.
Scaling writes across nodes using PL/Proxy works.
> Of course, I am assuming the disk system would be RAID 1, RAID 10,
> RAID 5, or RAID 6 for reliability purposes and that it is sufficiently
> redundant that you don't have to worry about an outage of your storage
> system.
The idea that you'll have a system that needs better write scalability
that isn't limited by the storage system is an unusual one, not the
expected case. And the trend everywhere in the industry is away from
giant redundant systems, and toward having multiple cheaper redundant
copies of all the data instead. It's impossible to protect against
things like environmental failure at any single location. Once you've
accepted that you have to be able to replicate this beast too if you
want high availability, you're back at having a multi-node problem
again. This is why the most active work is on distributed designs that
start on that basis, rather than projects trying to build more scalable
monoliths.
--
Greg Smith 2ndQuadrant US greg@2ndQuadrant.com Baltimore, MD
PostgreSQL Training, Services and Support www.2ndQuadrant.us
"PostgreSQL 9.0 High Performance": http://www.2ndQuadrant.com/books