CirquedeSQLeil (4/29/2010)
Thanks Lynn.The part that gets me with these questions is if the hardware failure was enough to corrupt the database and the full backup - it would seem to me that you would be unable to perform a tail log backup. After all the failure had to be severe enough that it was able to corrupt files on different drives / luns. Thus I always have to step back and think a little more about it. It must not have been too severe of a hardware failure because the database was able to be brought back online.
Actually, the basis of this question was not from a real world failure, but based on several posts here on SSC where people were truncating the t-log just prior to a full backup or simply truncating the t-log as a part of normal processing. I wanted to show why you don't necessarily want to to do things like that.
In this particular case, how would you recover to 9:00 AM Wednesday if you truncated the t-log just before the Tuesday night full backup and later discovered that the backup file was corrupt?