There are two usages we consider for that:
1 .Create a backlog of statements that I may need to execute on the slaves
databases in case of a fail-over in order to make it identical to the master
2. Control the connection pool so that in case the application wants
something that is not yet in the slave it will have to query the master.
On Wed, Dec 1, 2010 at 2:17 PM, Johan De Meersman <vegivamp@stripped>wrote:
> On Wed, Dec 1, 2010 at 7:01 PM, Frank <softwareengineer99@stripped>wrote:
>> Hmmm, why are you trying to verify whether transaction reached a slave
>> using COMMIT within bin logs? Why not just verify using the tables?
>> Care to clarify?
> I guess that's possible, but you'd need unique update sequences for every
> single column of every single table - quite a bit of overhead compared to,
> say, a general revision number a la version control.
> Bier met grenadyn
> Is als mosterd by den wyn
> Sy die't drinkt, is eene kwezel
> Hy die't drinkt, is ras een ezel