[rdiff-backup-users] Robustness to errors during backup

[Eric Jensen]

Related to the recent post on how to recover from a crash during a
backup, here's a question that has been on my mind lately.  Is there any
way to make rdiff-backup more robust so that it doesn't crash during a
backup?  I've been trying to get it to work with an SMB-mounted
filesystem, and every time I think I'm getting close, a different
filesystem error causes the initial backup to fail.  For example, I
found out the hard way that the filesystem I'm backing up to won't
accept files larger than 2 GB, won't allow creation of autorun.inf or
*.pif file (due to antivirus software), etc.

Now, these are limitations of my fileserver, so they're not a problem
with rdiff-backup.  The problem is that every time the running
rdiff-backup encounters one of these limitations (a file it can't create
for some reason), it brings down the whole backup.  When backing up a
100 GB+ filesystem, this is a big problem, as the process can run for
hours on the initial backup before crashing one one particular file.
I've also had rdiff-backup processes fail due to "busy text file"
errors.

Certainly any sort of backup software needs to report errors, but it
seems that ideally the software would continue with the backup even if
it encounters an error on a particular file.  I know nothing about error
trapping in Python, so maybe this isn't possible.  Failing that, is
there any setting I can use that would make rdiff-backup more likely to
continue after an error?

Thanks in advance for any thoughts.

Eric