Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Conflict resolution in Symmetric replication

Re: Conflict resolution in Symmetric replication

From: <cntc_at_my-dejanews.com>
Date: Thu, 30 Jul 1998 17:42:43 GMT
Message-ID: <6pqbaj$dt6$1@nnrp1.dejanews.com>


In article <6ppbb1$63o$1_at_nnrp1.dejanews.com>,   josephsha_at_my-dejanews.com wrote:
> hi,
>
> Could anyone give me more insight on the following:
>
> We have multiple databases spanned across different countries. We want to
> provide the update anyware feature for these databases, using replication
> mechanism(either Oracle or inhouse-developed). I am aware of the fact that
> conflict-resolution rules will be applied to resolve conflicts while
> propagating changes across the databases. But I am not very clear about how
> these conflict resolution works. For eg. consider the following scenario:-
> Suppose we have setup replication to happen every 1 hr across databases D1,
> D2 and D3. And assume that the conflict resolution rule is that the latest
> change is accepted. At 6:55 Row r1 in table T1 is deleted by user U1 in
> database D1. At 7pm Replicator pulls the transaction log from all databases.
> At 7:01(Before replicator propagate the changes to all databases) Row r1 in
> table T1 is updated by user U2 in database D2. Now replicator propogate the
> changes (as of 7pm) to all the databases. So the change done by user U2 will
> be lost and user U2 will suddenly see the row r1 missing eventhough he has
> done an update.
>
> Could anyone clarify the above. Or if you think that my doubts are due to lack
> of knowledge in
> conflict-resolution-rules, can you tellme where to find more info on
> conflict-resoultion-rules in symmetric replication.
>
> Thanks & Regards,
>
>Joseph: Our company specializes in data replication. We have developed a

product that supports the functionality you have described. Here's how we address this issue. PeerDirect does not use transaction logs.......instead, we track the changes to the data between replication cycles and compare/distribute these changes based on the absolute time the change took place. The default rule in PeerDirect is "last update wins," however deletions always take precedent over updates in order to make them durable and reflect what the result would be if all users were connected to a centralized database. To learn more, please read our collision handling white paper located in the products section on http://www.peerdirect.com Regards Dave Bell email dbell_at_cntc.com

-----== Posted via Deja News, The Leader in Internet Discussion ==----- http://www.dejanews.com/rg_mkgrp.xp Create Your Own Free Member Forum Received on Thu Jul 30 1998 - 12:42:43 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US