GilaMonster (10/23/2010)
Wildcat (10/22/2010)
If you issue a FULL backup C without COPY_ONLY between A and B, then, you break the previous log chain A to B.Right?
Wrong.
Full backups do not and never have broken the log chain.
It's trivial to test and, if you want, I can find a blog post from the former program manager of the Storage Engine stating and proving this.
The only things that break the log chain are:
* Switch to simple recovery
* Backup log with nolog/truncate only
* Deleting a log backup (won't stop you taking log backups, will stop you recovering with them)
According to Microsoft, you should backup log after the Full backup C to establish a new log chain.
Link please.
I need to write a scathing email to either the author or the MS documentation people if such an official recommendation exists
How about this scenario? You have a system setup with a regular schedule, full backups daily and transaction log backups every 1h. One day a developer decides to take a manual backup not using COPY_ONLY, and for some reason decides not to keep the backup file and deletes it. The following transaction logs will be based on the backup the developer made, and unfortunately the server crashes the same day, before the next scheduled full backup happened. Would it be possible to restore the data using the previous full backup if the transaction logs between the previous full backup and the time of the crash, is kept?