[rdiff-backup-users] restore fails

[Travis]

On my system, I have a ram disk (tmpfs) set up for a disk-intensive
program, and I have rdiff-backup set to back up a snapshot from this
ram disk to persistent storage on the hard drive every hour.  Early
this morning, the ram-disk filled up and so my application started
doing all sorts of indeterminable things with the data in the
ram-disk, while the cronjob using rdiff-backup continued to back up
the corrupted data every hour.  I didn't think this would be a
problem, I just unmounted the ram-disk and remounted it with more
space, but now rdiff-backup fails to restore my files, it gives me
loads of errors like this:

Error reading /path/to/archive/server.log, substituting empty file.
Warning: Hash da39a3ee5e6b4b0d3255bfef95601890afd80709 of server.log
doesn't match recorded hash edca93ea92f5ac446256c6a733e9b2d65d8c5f68!
Error reading /path/to/archive/server.log.lck, substituting empty file.

Note that the rdiff-backup archive location has always had plenty of
space, and has never been manipulated with any tool other than
rdiff-backup.

Any ideas about what is going on or how to recover the data?

Thanks much.