We are contemplating some migrations to PostgreSQL, but I am having a hard time figuring out how to use the Continuous Archiving feature on the instance for our databases.
We would like to have, say a database for each application , in our cluster/instance.
And use continuous archiving to be able to maximize our chance of not loosing any data... And do Point In Tiem recovery of any of the databases.
So : one instance/cluster with many databases, and the ability to
restore any of them to any point in time, just like "some other RDBMS"
However - when reading about the WAL, it seems they are covering *all* databases in our cluster/instance ?
And then it is only possible to do PIT recovery/restore on all databases - so all applications will be nuked back to whatever time we restore to ?
Any suggestions on how to accommodate the wish that each database should be able to do PIT recovery ?
Do we need to create only one single database in it's own instance/cluster and then do WAL backup/continous archiving on that - or ?
Or should we use pg_dump every 15 minutes and say they will loose max 15minutes data if a single database must be restored... ?
Or have I completely mixed up Oracle/ PostgreSQL terminology :) ?
pgsql-novice by date
|Next:||From: Dem Fra Hedehusene i Danmark||Date: 2011-03-11 09:20:33|
|Subject: PIT Recovery of only certain database in the "cluster" ?|
|Previous:||From: Vibhor Kumar||Date: 2011-03-11 08:39:09|
|Subject: Re: regarding functions and storedprocedures|