Re: Batch replication ordering (was Re: [GENERAL] 32/64-bit transaction IDs?)

On Friday April 11 2003 2:38, Richard Huxton wrote:
> Surely the batching should be happening at the "master" end? That's the
> only place with the context to mark a "determinate state". Batch things
> as fine as you like at that end, and just have the slave process multiple
> batches if it can/wants to.

Huh?  I don't understand this comment.  How to retrieve properly ordered but
limited batches of transactions and process them on the slave is the
question.

Ed