[Duplicity-talk] Faster queries for old backups?

[Ben Escoto]

Hi, right now old duplicity archives can be somewhat difficult to look
through because listing which files have changed or which files an old
archive contains requires looking through the whole archive.  This can
obviously be inconvenient, especially over a slow link.

    So, it seems that duplicity, besides writing archives like it does
now, should also write a list of files to the remote end.  That way
queries could be done by just fetching the list, and not the whole
archive.

    But it might help if I had some input:  In general what would you
like to know quickly about older backups?  How about this system:
For a full backup backup, duplicity writes a metadata information file
(just like the one rdiff-backup writes).  This isn't used for backing
up, but just for queries about what is in the archive.  For
incremental archives, duplicity can just write the differences old and
new metadata.  That way asking what has changed between two backups
would just require downloading the diffs, which should be pretty
small.


-- 
Ben Escoto

pgpYtKKEHKNwo.pgp
Description: PGP signature