Distributed Databases Advice
Hi guys,
I have the following challenge at hand.
Oracle 8.1.6
Situation : I have a central database and 5 other regional database that get a small percentage of data from the central database. We replicate a total of 67 snapshots. Some are fast refresh others cannot be since they have complex queries associated with them.
Replication takes place every 2 hours.
Problem with current setup : We like to make the replication intervals as close to real time as possible. Secondly i seem to have a problem with locks on the mlog$ tables when more than one site tries to replicate the same data at the same time. Thirdly replicating to 5 sites puts a strain on the performance of the central database.
Any thoughts to help me with my challenge
By the way is my thinking right to say that snapshot replication is a pull and advanced replication is a push. ????
Kind Regards
Vishaal
Re: Distributed Databases Advice
Quote:
Originally posted by Vishaalc
Some are fast refresh others cannot be since they have complex queries associated with them.
If you want 'almost real time' data, maybe you should rethink these complex queries.
That's what I did, I have fast refresh on all tables in a under 2 minutes period. Because of not using complex queries I have 2 tables with more data than I need, but in my environment this is not a problem (of security or integrity). Even with excessive data being refreshed, it's faster, it's putting less overhead on network and less work on master database than it would with full refresh.