duplicity-team team mailing list archive

[Michael Terry <mike@xxxxxxxxxxx>]

So as I speculate in a recent merge [1], I think bug
1155345<https://code.launchpad.net/bugs/1155345>,
bug 720525 <https://code.launchpad.net/bugs/720525>, and bug
662442<https://code.launchpad.net/bugs/662442>may all have the same
underlying cause: a missing diff or snapshot entry in
the chain.

I also imagine that this is due to a bug when restarting a backup, just
because that seems likely with this sort of issue.

Maybe a file that has a diff entry in sigtar.part has its actual diff
contents in the vol file deleted when resuming?  (We do delete the last
volume when resuming in case it didn't finish.)  But I wasn't able to
create the situation manually.

I just figured I'd ask the brain trust if they could brainstorm any ways
this situation could occur...  It would be lovely to kill all these
long-standing bugs in one fell swoop.

[1]
https://code.launchpad.net/~mterry/duplicity/catch-seq-copy-error/+merge/186106

-mt