yoda wrote:
> I'm looking for information regarding multimaster replication in Oracle
> 10g. I've found out that asynchronous replication would suit our needs
> well. The exception is the data conflict that occurs when two users at
> the same time creates a new record in the exact same table in their
> local database. For this I'm thinking if it is possible to request
> propagation of queued data from a trigger that is fired off on when a
> new post is created? I'm thin that this could eliminate or at least
> largely minimize the problem of unique key violations on creation of
> new records in the database. Is it possible to, besides scheduled
> propagation, also request propagation from a trigger?
>
> Thankful for any answers.
>
> /yoda
>
Your users must be well synchronized, then, or you design has a flaw.
I suspect the latter, as replication uses the primary key to
distinguish records from each other.
And yes, you can refresh on commit (near real time) through on demand,
and every job timing you can think of in between.
But I stick to my suspicion about your design.
--
Regards,
Frank van Bortel
Top-posting is one way to shut me up...
Received on Thu Mar 16 2006 - 13:52:47 CST
