Mark Fowler wrote:
So,
It's nearly the new year and the time for new projects. On my project
radar for next year is "consider (better) sharding our MySQL database."
Standard consultant question: Why?
"go faster/scale up" is not a good reason.
If you just wanted to go faster you would have moved to Pg or similar :-)
There is no magic solution - the DB and application and usage should be
reviewed to see if some minor tweaks can make major improvements.
Picking a solution without first knowing what the existing problems and
bottlenecks are smells of "I wanna play" :-)
Jacqui
p.s. I have wanted to play with a Pg SSI over a OpenMosix image for
ages but have yet to make the time. The original question sounds
similar to the reasons for my wanting to play with Pg over OpenMosix
(because it sounds like fun and May help scale up a app) - I had a
good excuse a year or two ago but today with >16+ CPU servers, Hadoop
etc there are better things to do with my time.