The rsync utility is a mandatory tool for your command-line work, and you need to learn how to use it for easy, safe, quick laptop-desktop synchronization. This article introduces rsync and mentions several related tools, all of which provide easy synchronization procedures.
…for both Windows and OSX that use RSYNC…
DeltaCopy (Windows) – http://www.aboutmyip.com/AboutMyXApp/DeltaCopy.jsp
Carbon Copy Cloner (OSX) – http://www.bombich.com/software/ccc.html
Mike Bombich’s site also includes a tutorial on doing remote backups of OSX using RSYNC.
Carbon Copy Cloner uses rsync? That’s cool – I’ve used it for 2 or 3 years and didn’t realize that.
For laptop-desktop synchronization where you might change different things on each computer, I like Unison:
http://www.cis.upenn.edu/~bcpierce/unison/
(It’s cross-platform, open source, with Windows, gtk2, and Mac OS X GUIs.)
Unison is an excellent program; I simply couldn’t be without it. Rsync is great for one-way stuff like backups, but when you are working on two machines regularly (not just one of them) then Unison comes into its own and is extremely helpful.
If you use a central server as one of the endpoints in the Unison synchronisation, you can synchronise >2 machines.
For backing up my laptop to my external hard disk, haven’t found a better tool than rsnapshot:
http://rsnapshot.org/
I just run it in my cron. Unlike the other rsync-related tools the author mentions, it is mature, well-maintained and gives access to all the underlying rsync options if you so choose.
Edited 2009-04-16 04:01 UTC
If you are interested in using automated (i.e : passwordless) rsync backup while keeping it fairly secured, you might be interested in the following :
http://www.sakana.fr/blog/2008/05/07/securing-automated-rsync-over-…
so that you can make your own 1-click backup or, even better, schedule it through cron (unix/linux) or Scheduled Tasks (Windows).
Of course you can’t use rsync alone to do your backups (for one thing, you once synchronized, you won’t be able to have a previous version of a file, for example if it got corrupted).
The following article provides a way to get this feature and some more by using rsync + a Solaris 10 machine as the backup server :
http://www.sakana.fr/blog/2008/07/05/backups-a-personnal-implementa…
It is easy to set up rsnapshot (see my comment above) to use ssh transport.
A page specifically about this topic:
http://troy.jdmz.net/rsnapshot/
Using rsnapshot is much more convenient than using rsync alone.
As someone else pointed out rsync alone is usually not good enough. The really interesting solution is when you combine rsync with open solaris zfs. For more details see http://milek.blogspot.com/2009/02/backup-tool.html
Edited 2009-04-16 10:08 UTC
http://gadmintools.flippedweb.com/index.php?option=com_content&task…
Gadmin-rsync looks nice too. I haven’t tried it yet but looks like it could be a decent GUI rsync tool.
I’m going to guess that the system clocks on the two systems should be set correctly. Most people take NTP for granted, but I’ve run into more than a few problems with Linux being unable to update the hardware clock and that causing various problems.
The problem I have with rsync is that it doesn’t preserve all file attributes when pulling a backup from a linux to a windows machine.
Well, creating a tar archive on the linux machine and just pulling that one over rsync helps, but that’s not a very good solution, any suggestions?
rsnapshot, pci-sata adaptor, for some reason
seems quicker than ide (master ) > ide (slave)
with rsnapshot.
………..
the post above I am replying to though, “windows”,
you may consider the shareware Bootit. I use it
for image backup (it does BATCH!) of fat32 and
freebsd partitions. I also use it to format
a FS to “freebsd” (it is in the dropdown menu if
one knows beforehand) before installing so so when
the somewhat arcane freebsd installer operates, I know
precisely where to install. Slightly costly (35$) but
way easy. Initially bought it in the 1990’s when
knowledgable persons raved about it in Usenet.
Well, duh! Only a single device can be active on an IDE channel at a time. So you read data from the master to a buffer, then write to the slave. It’s a half-duplex operation. It’ll always be faster to use two separate IDE channels (read from primary master, stream directly to secondary master), or any two separate I/O channels.