On Fri, Jul 13, 2012 at 3:07 PM, Bryan Keadle (.net) <bkea...@keadle.net> wrote:
>
> I just recently been introduced to BackupPC and I've been pursuing this as a
> VM appliance to backup non-critical targets. Thus, as a VM, I would use
> remote storage (iSCSI) to provide capacity instead of local virtual disk.
That can work, but given the ease of apt-get or yum installs on any
linux system, what's the attraction of a VM? You'll get a lot of
overhead for not much gain. And if you end up sharing physical media
with the source targets, you shouldn't even call it a backup.
> Les - as for the offsite copies of the archive, you're speaking of a backup
> of the backup. For the purpose that BackupPC provides, non-critical data,
> I'm not so concerned with backing up the backup.
What's your plan for a building disaster? If it is 'collect the
insurance and retire' then you probably don't care about offsite
copies...
> However, should BackupPC
> start holding backups that I would need redundancy for, what do you
> recommend? What are you doing? Since I'm thinking SAN-based storage for
> BackupPC, I figured I would just use SAN-based replication.
If you aren't sharing the media with the data you are trying to
protect, that would work. But, if you have the site-to-site
bandwidth for that, it would be much cheaper to just run the backups
over rsync from a backuppc server at the opposite site. I have
mostly converted to that approach now, but an older setup that is
still running has a 3-member RAID1 where one of the drives is swapped
out and re-synced weekly. These were initially full sized 750 Gig
drives, but I'm using a laptop size WD (BLACK - the BLUE version is
too slow...) for the offsite members now. Other people do something
similar with LVM mirrors or snapshot image copies.
> But should that
> not be wise or available, would you just stand up an rSync target and just
> rSync /var/lib/BackupPC to some offisite target?
The number of hardlinks in a typical backuppc archive make that a
problem or impossible at some point because rsync has to track and
reproduce them by inode number, keeping the whole table in memory.
I think a zfs snapshot with incremental send/receive might work, but
you'd need freebsd or solaris instead of linux for that.
--
Les Mikesell
lesmikes...@gmail.com
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
BackupPC-users mailing list
BackupPC-users@lists.sourceforge.net
List: https://lists.sourceforge.net/lists/listinfo/backuppc-users
Wiki: http://backuppc.wiki.sourceforge.net
Project: http://backuppc.sourceforge.net/
