Checking health of standby database

-----BEGIN PGP SIGNED MESSAGE-----

Hash: SHA1

Twice in the last year I've needed to activate my standby database only to find out that even though the logs are being applied correctly, there is something wrong with it. In one case a corrupt log was shipped to it. In the other there was datafile corruption that went completely undetected. There was also missing datafiles that also went completely undetected (i.e. no failures to apply logs and nothing in the alert log). Both times the fact that the standby could not be activated resulted in data loss and many hours of downtime on a critical database.

I want to implement a script to regularly check the health of the standby databases so I'll know if any problems exist that would prevent it from being activated. My initial thought was to open it read-only and do an export to /dev/null, but that's taking too long. The database is 75g and I had to kill the null export after 15 hours on one table. It's not hung. I've queried v$session_event and the export session is running. The export is just taking that long.

I have a 6 hour window at night in which to complete the health check. Can anyone suggest a better alternative that would work in read-only mode? (Analyzing the objects won't work because it will try to write to the data dictionary.)

-----BEGIN PGP SIGNATURE-----

Version: PGPfreeware 7.0.3 for non-commercial use <http://www.pgp.com>

iQA/AwUBPHzwZGm5A4kkb4ZsEQIbdACgqeqEx6ygBKGhFDZggS60EUqP8eMAmgNx N9VMgUj2bpvmvFxP73DKW1Vg
=Ebft
-----END PGP SIGNATURE-----
Received on Wed Feb 27 2002 - 08:42:47 CST