[Duplicity-talk] retries / time-out

[Dan Muresan]

Continuing my quest to back up to a DAV-enabled host, I ran into a
different problem: the initial "full" backup fails every time with a
Python socket time-out error. Issues:

* is there a way to increase the timeout? (hopefully it is a per-file
timeout)

* the default value of 5 for --num-retries needs to be documented

* is --num-retries per-file, or do failures accumulate across all files?
The retry quota should be per-file, otherwise large backups will never
be robust.

* it seems to be easier to use something like cadaver to upload the
initial full backup (about 100 volumes). Will such an uploaded
repository be valid?

In particular, are the timestamps of individual volumes significant? I'm
referring to the timestamps from uploaded volumes' metadata, as set by
the DAV server, not the timestamps built into volume names when
--short-filenames is inactive.


Thanks,
Dan