Just throwing out an idea here: If we are going with a database or some sort of keystore for the metadata. We can still generate one "keystore" file for each revision. But the current mirror metadata could contain the full history.=Making sure we could always go back to the previous keystore and recover manually...
On Sat, Jun 6, 2020 at 9:27 AM Robert Nichols <rnicholsnos...@comcast.net> wrote: > On 6/5/20 4:44 PM, Arrigo Marchiori wrote: > > If we were going to substitute a lot of files with a single file (that > > is what a SQLite database is in the end, right?) then we may somehow > > introduce a "single point of failure" for the whole backup. > > The rdiff-backup archive structure is already a rather fragile ad-hoc > database with single points of failure. If you lose the current > mirror_metadata file, the entire history is unrecoverable and you > are left with just the current mirror and whatever file attributes > are compatible with the filesystem of the archive. > > -- > Bob Nichols "NOSPAM" is really part of my email address. > Do NOT delete it. > > > -- Patrik Dufresne <ikus...@gmail.com>
