Hi Paul: On 5/9/09 5:53 AM PDT, "Paul McCullagh" <[email protected]> wrote:
> Hi Robert, > > On May 8, 2009, at 6:36 PM, Robert Hodges wrote: > >> I think even the slave restart idea has problems, because on restart >> you can >> violate serialization requirements on the start transaction. >> > > As far as I know, the start time of a transaction is not significant. > > Or do you maybe have an example of where this is the case? > > The start order just helps us to start the transaction on the slave at > more or less the correct time (to ensure the correct commit order). > > Which means as long as the commit order is maintained, transactions > can be restarted at any time. > I was having a flashback to statement replication where start time is significant for any update statement that has a where clause. Drizzle replication will be row-based so this does not seem to be a problem. Cheers, Robert _______________________________________________ Mailing list: https://launchpad.net/~drizzle-discuss Post to : [email protected] Unsubscribe : https://launchpad.net/~drizzle-discuss More help : https://help.launchpad.net/ListHelp
