I'm a very new user of Amanda and we would replace our old rdiff-backup for some reasons, including the incoming use of storage clusters and S3.
I did some tests but there is something very unclear for me even after searching in this forum and in Google.
Our use of Amanda would be this one :
We have some file storage servers and we would make a daily backup of the data. Only one directory /mnt, we do not mind of the OS, it's not a problem. Not a full backup each day, they are in each server some TB of data, so a full backup each month would be an otpion.
If one of them crashes, we would restore it exactly at it's state at the last daily backup. We will never use Amanda to restore a file or directory deleted by accident, that type of backup is managed by other applications handling ( very ) long retention periods.
As an exemple, Monday, two files are created 1.txt and 2.txt. On Wensday, 1.txt is deleted. On Saturday, the server fails. We would to find in the restore the file 2.txt and not the files 1.txt.
It is a very KISS use , we want to restore a server how it was at the last daily backup before the crash, no more, no less.
Right now, all I'm able to do, it's to restore 1.txt and 2.txt.
To erase manually deleted files and directories is not an option, we simply do not know what was deleted and we will never know it !
How can I do that ?
Thank for your help.