Re: [rdiff-backup-users] rdiff-backup for email

[Michael Simms]

Hi,

All good points below, thanks!

We use UW-Imapd with mbx mailboxes. That means a user's inbox is 1 file. 
Users can also store other mail folders on the server, but it's not like 
maildir where every message is a new file. This can lead to problems when 
a file changes while it's being backed up, but that happens in our current 
backup environment also.

We currently use Legato and do only full backups. That's because, due to 
the way the mail is stored, a lot of files change every day. Incrementals 
were using 70% of the storage, and taking longer than fulls because of the 
extra processing time. I see rdiff-backup as a way to cut down 
significantly on the amount of space being used for incrementals. I'll 
have to weigh the pros and cons.

--
Thanks,

Michael


On Sat, 11 Nov 2006, roland wrote:

> Hi !
>
> how many files ?
> what type of files? (many small ones, some large ones like database backing 
> store...)
> what backup stragegy?  (1xfull, 6xincremental in a week? )
> could you perhaps describe your current backup strategy ?
>
> it´s important to put backups with mission critical data to offline media.
> harddisks and a "write backup to harddisk"-tool alone may be not safe enough.
> also important is to do recover tests at some interval.
>
> furthermore, you have no guaranteed support with rdiff-backup, i.e. if you 
> get a real problem,  there is no warranty that you get proper help.
> unfortunately the main developer is quite inactive and i have seen the one or 
> other questions/problems on this list being unresolved due to lack of people 
> helping or lack of knowledge about rdiff-backup internals.
>
> the positive thing is:
> rdiff-backup and especially the backup format is "understandable" - i.e. you 
> can investigate into it yourself if you have some programming skills.
> the latest backup is always a 1:1 copy of your original data (like rsync), 
> which can be recovered without rdiff-backup working at all.
> furthermore, you can restore older files without rdiff-backup if you have 
> understood how files being stored (reverse diffs) , so you can recover single 
> files from the backup repository with nothing more than rdiff utility.
>
> rdiff-backup is great, no question, but you should think twice and do lot`s 
> of testing before, especially run-time tests.
> rdiff-backup's "processing overhead" for each file is significant, so you may 
> get into "timeslot trouble" if you have millions of small files.....
>
> regards
> roland
>
>
> ----- Original Message ----- From: "Michael Simms" 
> <address@hidden>
> To: <address@hidden>
> Sent: Friday, November 10, 2006 10:53 PM
> Subject: [rdiff-backup-users] rdiff-backup for email
>
>
> > Hello,
> >
> > I'm new to the list, and to rdiff-backup. We're looking at alternatives for 
> > backing up email. It's currently 3TB, and growing quickly. Is anyone out 
> > there backing up email, or other 'mission critical' data with rdiff-backup?
> >
> > --
> > Thanks,
> >
> > Michael Simms
> > Computing and Networking Services
> > University of Toronto
> >
> >
> >
> >
> > _______________________________________________
> > rdiff-backup-users mailing list at address@hidden
> > http://lists.nongnu.org/mailman/listinfo/rdiff-backup-users
> > Wiki URL: 
> > http://rdiff-backup.solutionsfirst.com.au/index.php/RdiffBackupWiki 

--
Michael