DimitriK's (dim) Weblog

MySQL Performance: Full Time since Today ;-)

Until now, all my work around MySQL Performance was just a part of fun rather my main job... But things are changing since today ;-) I'm moving to MySQL Team, and will work on MySQL Performance full time from now on! ;-) Which is meaning yet more fun (I hope ;-)), more blog posts (I'm sure ;-)), more benchmarks (of course), and more brain work (even more than sure ;-))..

And as I've mentioned benchmarks - many of you complained over a time that we're not really testing what we should.. Well, everything is relative and depending directly on what you're trying to fix or analyze.. ;-) So, I'll not stop running dbSTRESS or Sysbench workloads (or TPCC-like, DBT-2, etc.).. But, I'm also looking for other various test scenarios to reproduce different MySQL performance problems you're observing in your production, or your customer's sites, etc. Please, leave comments on this post, or contact me directly if you prefer. I'll try then to group all such test scenarios together and then ship them as extension to dbSTRESS or any other way (I don't have yet a final test kit, but many ideas are running in my head ;-)), publish performance results and their analyzes (as usual).

As well if you have in mind some top priority performance problems in MySQL 5.5/5.6 - please share them too. I promise you to add them all into my TODO list, and once the impact is validated - see how better it should be fixed (but, of course, I have to be able to reproduce it first ;-)) I'm already have in head issues with adaptive flushing, purge lag delay, Performance Schema overhead, etc. - but you may see other things and have other priorities..

So don't wait - share! ;-)

Rgds,
-Dimitri