, , , , , , ,

This is a sudden thought that starts biting me…….

Let us consider a scenario, where we have a big application with thousands of user, using Oracle Database, and the application do not support the logical standby database. So for a DR setup, we are using a combination of primary database at main Data location and a physical standby database at the Disaster Recovery location using Oracle Dataguard technology.

And let us also consider the case, where the hardware used in secondary standby database is of entirely different hardware, I mean much lower in capacity (probably the ideal case – people may not want to spend as much money as it is for main data center and keep it idle – usage is much less unless there is a disaster or drill).

Now at the normal condition, any changes happening at Primary database will be applied using the Archived Logs in the standby database, and the statistics of the database objects in the primary database. Now when primary fails over to secondary, secondary database will have statistics which is actually of no use, because it is the statistics generated relevant to the hardware at primary site.

So expected is the slow response (At least the queries and reports) which will be much slower, when DBA starts generating the statistics…..