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

[Jim C. Nasby]

On Tue, Jun 26, 2007 at 11:32:06AM -0500, Charles Duffy wrote:
> Xavier Bertou wrote:
> > I must agree with what Frederik said. rdiff-backup works just fine
> > when rsync works just fine, ie when you don't need a backup except for
> > stupid user deletion of files. When real bad things happen,
> > rdiff-backup doesn't handle any errors. File I/O error? The whole
> > backup fails, and I get zillions of errors on checking the
> > incrementals. I had at some time a very bad filesystem issue which
> > scared the kernel whenever I was trying to access a file, Linux just
> > killing the process. Of course this never had happened to me in the
> > last 10 years of Unix administration, but it happened a few months
> > ago, and again, no backup for a few days until I figured it out, and
> > no way to recover the incrementals.
> 
> I have a system of cron jobs and wrapper scripts surrounding my
> rdiff-backup installation which kills dead processes, reports on which
> backups succeeded and failed, ensures sure that --check-destination-dirs
> is run on any folder that needs it before any new backup is started, and
> disallows multiple connections attempting to backup from or restore to
> the same host at once.
> 
> rdiff-backup is indeed not quite as robust as it could be
> out-of-the-box, and it does indeed seem to trigger some issues in ext3,
> but with appropriate surrounding infrastructure it WORKSFORME.

Of course this begs the question: why isn't that stuff built in to
rdiff?

The big thing I'm missing as compared to rsync is the ability to deal
with unreliable links. I normally won't backup my laptop when I'm on the
road because it's a royal PITA anytime a backup fails partway though.
Granted, I've been lucky enough that that's never taken any manual
intervention once I got everything setup properly to get an initial
backup, but it still takes a long time to roll everything back. And
really I'd rather it not just roll everything back, since that means I
just have to re-do all the work that the failed backup had already
accomplished.
-- 
Decibel!, aka Jim C. Nasby, Database Architect  address@hidden 
Give your computer some brain candy! www.distributed.net Team #1828

pgpRP0UOHnC1N.pgp
Description: PGP signature