On Wed, 2010-02-03 at 00:31 +0100, Rene Veerman wrote: > On Wed, Feb 3, 2010 at 12:18 AM, Ashley Sheridan > <ash@xxxxxxxxxxxxxxxxxxxx>wrote: > > > The problem is where 2 people choose the same instant to perform an > > action on your site that inserts a record into your db. The db engine > > inserts them one after the other, and then responds about the max(id) to > > your PHP script. Then, you now have 2 people who have the same max(id) > > retrieved, but one of the values is wrong. > > > > well, i only use getmaxid()s for inserts. > the timelag between getmaxid() and the insert is so small it'd take 300-800 > insert-requests/sec > (on that particular table) before an error condition would arise. > in which case, a tested piece of sql would fail, and can be routed through > the retry functions . > These would imo provide ample timing re-randomization, aswell as a measure > of stress-relief for both php and mysql server. > You may correct me if i'm wrong :) > > BTW: php core developers: can we have a sleep() that accepts a float? :) It's the reason transactions exist, to prevent things happening like this. When you have two actions where one is dependent on the other, unless you have a way to tie them together so that they can't be broken, you run the risk of collisions. Thanks, Ash http://www.ashleysheridan.co.uk
