Re: [rdiff-backup-users] Resuming initial backups

[Josh Nisly]

Maarten Bezemer wrote:
> Hi,
>
> On Fri, 22 Aug 2008, Josh Nisly wrote:
>
>   
> > My question is this: couldn't rdiff-backup detect this and handle it 
> > automatically? I've added code to my private branch that checks if there 
> > is no current mirror and one or fewer error_log and mirror_metadata 
> > files. If so, it removes them and runs normally. Does this seem like a 
> > reasonable way to handle it? If so, I'd be happy to submit a patch 
> > implementing it.
> >     
>
> Initial backups are (in my case at least) attended rdiff-backup runs. If
> the first run is aborted before it finished (e.g. because of network
> outage), I can restart it by hand; this is a well documented approach,
> also using the --force option.
> (Actually, my first backup always is a plain rsync, because that's much
> faster than rdiff-backup. So I already need the --force option to get the
> rsync-ed copy into the rdiff-backup meta tree.)
>   
Ah, that would be the difference. Typically, my end users download and 
run rdiff-backup themselves.
> Before submitting a patch, make sure that you handle things correctly
> under all circumstances. If your patch removes the need for the --force
> argument, it should be a safe operation. Never do unexpected or
> potentially unwanted things without the user explicitly using the --force
> argument.
>   
Indeed, that's the reason I raised the question here. My thinking is 
that if there is no current_mirror marker, and one or fewer error_log 
and mirror_metadata files, there cannot be any historical increments, so 
the operation cannot lose data.

Thanks,
JoshN