[Fed-Talk] Re: Backup 3
[Fed-Talk] Re: Backup 3
- Subject: [Fed-Talk] Re: Backup 3
- From: "Karan, Cem (Civ, ARL/CISD)" <email@hidden>
- Date: Tue, 11 Oct 2005 16:44:59 -0400
- Thread-topic: Backup 3
> You might be able to use rsync to do what you want; look at the man
> pages for more information on it. If you want to be able to
> recover old
> information (e.g., you have a file that you modify over time, and you
> want to keep old versions of the file around) look at subversion
> (http://subversion.tigris.org). That will allow you to maintain past
> information as well.
I should point out that rsync can be used to keep old versions of
files, if you use the right option you get one old version, if you
assign the date to a shell variable and use that as the backup label
you can keep as many old versions as you have disk space for.
In fact I use a series of rsync scripts to backup a collection of SGI
IRIX machines to an Apple G4 Xserve and 2 TB Xserve RAID (OS X 10.3
with case-sensitive file system) with precisely that technique. It
works automatically every night and I don't have 1 TB tarballs to
deal with. I rotate between the two 1 TB RAID sections manually;
however, even that could be automated. I also use rsync to maintain
a current copy of everyone's file structure on the G4 Xserve disks
(non-raided), that gives me a extra security for the current files
and the ability to NFS mount the Xserve's disks should the original
disk fail and people need to get work done.
If you need to squeeze space a bit further you can compress the older
versions with gzip after each script runs since you have the date in
a shell script. Probably bzip2 would even be better and I might have
even installed it.
In theory you could compress and uncompress the entire directory
structure that you're backing up into but I'm not sure the
practically of that when dealing with 1 TB of files and whatever
hardware you have.
Michael
-------------------
IIRC rsync will make a FULL copy, meaning that files that don't change
also get copied over. The advantage of subversion is that it only saves
the deltas. This can save a LOT of space. As an example, I have about
319 MB worth of project data stored in a subversion repository that has
481 different revisions. The repository is only 281 MB in size. If I
tried to save all those versions, using tarballs, it would require over
150 GB of space.
For those that are wondering if Subversion is doing compression, the
answer is 'I don't know'. I do know that subversion tends to put all
the deltas in one large file at a time, which means that fewer blocks
are wasted on disk for small files (which is what my project mainly
consists of). YMMV
Cem Karan
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Fed-talk mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden