Redo log size and standby database

Hi all,

I've recently taken over responsibility for administration of an OLTP database, about 15 Gbytes in size, running Oracle 8.1.7 on AIX 4.3.3. Generally, things look in pretty good shape, but one thing I am going to need to do in the near future is to create a standby database from this live one, to provide at least a read-only facility in the event of the live database becoming unavailable, and possibly to provide a full failover and activate the standby in the event of real trouble.

My question is this:

The redo logs on the live database are 100 Mbytes, in four groups, no multiplexing (just one member per group). Firstly, I'm going to mirror the logs and provide at least two members per group. However, I'm thinking of reducing the redo log size as log switches seem to happen only once per day on average. From what I know about standby databases, there is a chance of losing the current log if and when the live database becomes unavailable, depending on what causes it to become unavailable. In addition, if I remember correctly, there is a rule of thumb stating that log switches of a frequency around two per hour is a reasonable compromise between short recovery time and performance.

Does my reasoning sound, well, reasonable, or are there other factors to take into account?

Thanks for any advice and opinions.

Tim Kearsley
DBA, Milton Keynes Council Received on Tue Sep 03 2002 - 05:12:12 CDT