[rdiff-backup-users] How much extra space should there be on a target file system?

[Theodore Wynnychenko]

Hello

I have been searching for an answer or guidance to this question, but have
not been able to find anything, so I will ask here.

I am sorry if it has been asked/answered before.

In any case, I am planning on "updating" my backup hardware, and will
probably switch from rsync to rdiff (I have used rsync for many years, and
have been happy with it, except for that one time years ago when I deleted a
bunch of files and did not notice until about a day later, and 10 minutes
after the final backup had been synced - a bit disappointing).

Anyway, I am wondering if there is any guidance on how much larger the
target file system should be compared with the source file system when using
rdiff?

Clearly, rdiff stores more information, so, in theory, if you have two file
systems of exactly the same size, and the source is completely full, then
rdiff would have nowhere to store any historical changes.

Is there a rule-of-thumb, or some sort of calculation that will give
guidance on how much extra space there needs to be on the target file system
to reliably not have an instance when the target file system becomes full,
while the source still has space on it?

It seems this would be based on the frequency of updates, the size of the
file system, the number and size of files, the types of changes made to
those files, how long backups are maintained, and probably other things.

I am not looking for anything exact, more just a way to guesstimate.  But, I
am also not opposed to math either, if there is a calculation that can help
determine the target file system's appropriate size.

Thanks
Ted