rdiff-backup-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[rdiff-backup-users] rdiff-backup features?


From: Joe Beda
Subject: [rdiff-backup-users] rdiff-backup features?
Date: Mon, 9 Jul 2007 09:00:08 -0700

Hi everyone,

I'm new to using rdiff-backup.  I've been using Scott Ludwig's
Link-Backup (http://www.scottlu.com/Content/Link-Backup.html) but the
situation has changed and I'm looking for a new solution.  (I work
with Scott and he is in the same bind I am).   I have a linux machine
hosted in a datacenter and I do nightly backups to that machine over
my cable modem.  Most of what I'm backing up are image files.  I have
a small kid and like to go on photography trips so I sometimes have
large bursts of data (10GB or so at a time is not uncommon).  Because
of this, I usually can't fit in an entire backup without dropping a
connection (or stopping for regular daytime use of the cable modem).
In addition, it is not uncommon to want to rename and move files
around.  Obviously, I don't want to be moving lots of data around (and
wasting space on the server) for simple renames.

What has changed is that I've started using Lightroom to manage and
tag these photos and it writes to jpg files to change the metadata.
This causes link-backup to thrash.

It looks like rdiff-backup might be a good starting point, but I'd
need to add some features first before I could start using it.

Here is a summary of what link-backup is good at:
1) Backs up over multiple sessions.  If a backup fails, most of the
progress so far isn't lost.
2) Robust against file renames/moves (as long as the data hasn't changed).
3) Out of band "sneaker net" updates.  It is possible to load a bunch
of data on a laptop/external drive, go to the remote machine and move
the data over.
4) File system view of all previous versions.

Here is what I don't like about link-backup:
1) Doesn't handle diffs well
2) Uses hardlinks -- much harder to copy data from one volume to
another -- rsync works but it takes a long time and lots of memory to
solve all of the hardlinks.

To make rdiff-backup work for me, I'd have to add the following
features initially:
1) Support for recovering dropped or lost backups.
2) Support for handling moved files.
3) Add sneaker net type backups.

In addition, I'd be very interested in hearing anyone's experience
with the FUSE filesystem for accessing previous backups for
rdiff-backup 
(http://code.google.com/soc/2007/gentoo/appinfo.html?csaid=54C92C8CD707C04D).
This looks to be a Google SoC project and is very interesting.

So -- questions to the group: Are these features welcome?  Is anyone
already working on these?  Is there a development mailing list?

Thanks!

Joe Beda




reply via email to

[Prev in Thread] Current Thread [Next in Thread]