David Todd-242471 (2/24/2011)
Hi Folks,I've been thinking about this question, and looked through a fair bunch of Paul Randal's stuff around this issue and figure that the correct answer is actually missing: For me the correct answer is:
Maybe. Mabye you can recover to point-in-time, if there are no minimally logged entries in the log, or you are covered by a differential backup or whatever, but don't be an idiot and count on it.
Smart practice is not to see what you can get away with, but to make sure you do everything that you can not get away without.
Paul, in responding to the thought that now you have two full backups you can delete the oldest one, said something like: How do you know the current backup isn't corrupt? If it is corrupt then you need that backup you are about to delete!
Regards
David
The answer given was effectively "maybe, if there are not minimally logged entries in the log", but to go any further than that and add caveats about viable full and differential backups being available, no necessary backups being corrupted, lost or stolen etc would have over-complicated what was meant to be a simple question!
The point was not really to say "this is what you can get away with" but rather to say that if someone came to you in desperation with a database using the bulk-logged recovery model and wanting point-in-time recovery, you would not say "that's impossible" without checking their backups first.
Cheers,
Duncan