rescuing archive interrupted by "Too many network failures"

[David Prager Branner <brannerchinese@gmail.com>]

Sometimes a long and unattended back-up ends like this:

> tarsnap: Connection lost, waiting 30 seconds before reconnecting
> tarsnap: Connection lost, waiting 60 seconds before reconnecting
> tarsnap: Connection lost, waiting 90 seconds before reconnecting
> tarsnap: Connection lost, waiting 90 seconds before reconnecting
> tarsnap: Too many network failures

And on running tarsnap again with the --list-archives option I am aggrieved to find that no part of the archive in question has remained on the server, no matter how long tarsnap had been at work. This seems to me a terrific waste of bandwidth usage.

I'm wondering if there is not some way to preserve the portion of the data that has already been uploaded when this kind of error occurs. 

- dpb