On Thu, Aug 14, 2008 at 12:46 AM, Arjen Lentz <[EMAIL PROTECTED]> wrote: > Hi Giuseppe, Brian, all > > On a related note... if we're ditching MyISAM... if we can enforce that > engines are somewhat transactionally aware, then we can get rid of much of > MySQL's misbehaviour, generally related to having to make autonomous > assumptions. > Default values are part of this, as are automatic (potentially lossy) > casting, data truncation (string length, int range boundary), etc. > > The one thing that would need to be fixed is two-phase commit for multiple > engines. I believe most of the code for this exists but has some issues.
Would it be better to get rid of it? Build in some kind of simplified version of it like what used to exist for InnoDB in 4.x, which did not disable InnoDB's group commit, and assume/require that storage engines will support this more efficient method? Or was that just a bloody nightmare and I'm falling victim to good-old-days syndrome. The problem I'm pointing towards is that currently the synchronization between the binlogs and InnoDB causes a lot of extra fsyncs for each commit. _______________________________________________ Mailing list: https://launchpad.net/~drizzle-discuss Post to : [email protected] Unsubscribe : https://launchpad.net/~drizzle-discuss More help : https://help.launchpad.net/ListHelp
