Re: [rdiff-backup-users] ANN: rdiff-backup-fs 1.0.0

[Heiko Baumann]

i guess for most users its not a problem that the initial mount takes 
some time. one can do an automatic umount/mount after a new backup job 
is run.

just checking that the mountpoint is empty should be fine. in my case 
there was a file named ".keep" in the mountpoint (some gentoo stuff to 
prevent deletion of a directory on package uninstall) that forced the 
mount to fail. so you also should check for dot files. :)



On Mon, 14 Feb 2011 22:31:11 +0100, Filip Gruszczyński 
<address@hidden> wrote:
> > works much better than archfs which was to slow on bigger repos.
>
> Yeah, it even used to not work on bigger repos ;-)
>
> > one thing to mention is that i got the message "fuse: mountpoint is 
> > not
> > empty" _after_ some minutes of waiting for rdiff-backup-fs to do 
> > some work
> > (probably reading my repo), just to see that it refuses to mount my 
> > repo and
> > start it again with an empty mount point :)
>
> It is extracting metadata from the repo. This is another thing, that 
> I
> need to take care of, so the mount would be faster (though then
> entering some revision would take longer).
>
> > it would be great if you could do this check before any other time 
> > consuming
> > stuff.
>
> I believe simple checking contents of mount directory should be
> enough, right? If yes, I'll implement that.

> > and i have one question. is it safe to have a repo mounted with
> > rdiff-backup-fs while running an rdiff-backup to the same repo?
>
> I don't think that should break anything, but filesystem doesn't get
> updated and I can't guarantee it will continue to work. If you have
> periodical backups I suggest unmounting the filesystem and mounting 
> it
> again afterwards.