[geeks] rsnapshot backup revisited

[Charles Shannon Hendrix]

I tried rsnapshot for the last couple of days, and thanks to those who
recommended it.  

I got curious to know which files were changing, and/or why some hourly
backups were so huge.

This command will tell you where your file churn is:

% find <dir>.<gen> -links 1 -a -not -type l -a -not -type s

In other words, find all files that have only one link because they are unique
to the named backup. symlinks and sockets always get copied so you have to
ignore them.

This helped me find large files that changed, and save space in the
backups with sometimes minor changes in when those files get updated, or
maybe change how that data is stored.

For example, rsnapshot sucks for backing up mbox mail files, because the whole
file changes for every message added or changed. Switching to MH or maildir
folders would save you quite a bit of space.

In any case, it is also enlightening to see what files regularly change.

It also helps identify candidates for exclusion, like cache files and things
like that you really don't need backed up, and which can take up a lot of
space.

-- 
shannon "AT" widomaker.com -- ["Tara is grass, and behold how Troy lieth
low--And even the English, perchance their hour will come!"]