On Jul 3, 2009, at 8:01 AM, MARK CALLAGHAN wrote:
> I think that Drizzle has a (Drizzle only) fix for this in progress.
Yes, no FRM files :)
Though the solution is only perfect for an engine which can handle
recovery/manage its own data dictionary. Only Innodb and Archive can
do this at the moment via the design, but we expect that engines like
PBXT could as well.
I am hoping that the next version of the Innodb embedded engine will
give us a bit more reach into its internal data dictionary. That would
make all of this even simpler (though whether we will have a 1=1
schema matchup will be an issue).