Possibly moving to InnoDB

Max Kanat-Alexander mkanat at bugzilla.org
Sun Aug 6 15:39:36 UTC 2006


On Sun, 2006-08-06 at 17:16 +1000, Bradley Baetz wrote:
> >         But wait, that makes sense. You mean if a SELECT is blocking everybody
> > else from running, it kills the SELECT, right?
> 
> No. It kills the other one - ie the one that is not running.

	That's pretty silly.

> Yeah, but I've had it happen with concurrent update/delete. We 'fixed'
> it by changing delete to select first then delete if there was a match

	Odd. Thankfully we rarely have a concurrent UPDATE/DELETE. We
occasionally have some concurrent INSERT/DELETEs, though (like
logincookies).

> >         Okay. So we should recommend that if people are going to use
> > replication, they use at least 4.1.11 or 5.0.3.
> 
> It still doesn't help, though - it just makes it less likely, since by
> increasing the time before theres an error, you've decreased the number
> of times an error will occur.

	Yeah. I mean, I know that many huge sites use replication just fine on
InnoDB, including LiveJournal and Wikipedia, so some of it probably
depends on the structure of the application.

	-Max
-- 
http://www.everythingsolved.com/
Everything Solved: Competent, Friendly Bugzilla and Linux Services




More information about the developers mailing list