Hi!
On Jul 29, 2009, at 11:07 AM, Jonathon Hill wrote:
So are you doing row-level replication or something different
altogether?
Something closer to row-level, but I wouldn't use that word for it.
We transport changesets of a table based on events. DDL is still
statement based, but at some point in the near future this will change
as well (which will also clean up part of the parser believe it or
not...).
We replicate on a primary key/unique key predicate for each row and
provide "headers" for what the table definition. Takes up a little bit
more memory, but allows for a number of different replication patterns/
federated replication.
Cheers,
-Brian
_______________________________________________
Mailing list: https://launchpad.net/~drizzle-discuss
Post to : [email protected]
Unsubscribe : https://launchpad.net/~drizzle-discuss
More help : https://help.launchpad.net/ListHelp
