BackupPC-users

[BackupPC-users] My Solution for "Off-Site"

2011-07-10 02:53:21
Subject: [BackupPC-users] My Solution for "Off-Site"
From: Christian Völker <chrischan AT knebb DOT de>
To: "General list for user discussion, questions and support" <backuppc-users AT lists.sourceforge DOT net>
Date: Sun, 10 Jul 2011 08:22:25 +0200
Hi,
I just want to share my solution to keep an additional "backup" from the
original BackupPC store.

As we all know it's not really a good solution to rsync the BackupPC
datastore to somewhere else- due to the hardlinks. Doing manual image
copies (ie by swapping the drives of a RAID-1 array) has the big
disadvantage as it's a manual step.

So I decided to combine a couple of other techniques here:
First, my BackupPC is running as a virtual machine on VMware ESX host
sharing datastore and resources with the machines to back up. So the
obvious disadvantage is the case when the ESX host fails- how should I
restore this guy and the BackuPC machine? Well ESX is fairly stable but
you never know.
My storage uses in total 952GB of backup data. So it's really no good
idea to do an rsync here. Swapping drives manually is no good either as
the ESX host would complain.

So what I did was to set up a physical small sized box (old desktop
should work). No RAID involved. I installed there distributed remote
block device (drbd- use Google). Same on backuppc machine. So I have a
physical separated RAID1 available- just through network. Both drbd
devices are using LVM volumes as backing devices so I can enlarge/
shrink at will. The external server addditionaly uses the snaprotate.pl
script to create 4 snapshots of the drbd device at weekly rate.
The drbd device has BackuPC installed, too. So I can easily tell him to
take over and restore.

So with my setup I'm nearly prepared for everything at relatively low cost.
-backupc itself fails
+drbd one will take over after some minor (manual) steps.
-backupc wipes out it's storage (script failure or file system issue)
+I will roll back on the drbd to one of the previous LVM snapshots (up
to four weeks back)
-ESX host fails without removing backuppc
+Set drbd as primary and restore ESX (or just reinstall, it's faster)
-ESX host fails with wiping out the backuppc VM
+Set drbd as primary and restore everything from there on

So in summary I can easily keep my backuppc storage remotely in sync
with drbd and keep snaphshots to roll back weeks. The initial sync and
data migration to drbd device took 24hours while- you could reduce
backuppc downtime by not doing the dd command in parallel to the initial
sync.

Only disadvantage is the physical drbd is currently in same building (my
home) as the original ESX host. But this is not likely to change due to
my low external uplink bandwidth here- someone could use the drbd proxy
to use small lines for sync. But this proxy is not available as free
software. For me it's fine- if my house burns down I have more serious
issues than my BackupPC storage ;-)

GReetings

Christian




------------------------------------------------------------------------------
All of the data generated in your IT infrastructure is seriously valuable.
Why? It contains a definitive record of application performance, security 
threats, fraudulent activity, and more. Splunk takes this data and makes 
sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-d2d-c2
_______________________________________________
BackupPC-users mailing list
BackupPC-users AT lists.sourceforge DOT net
List:    https://lists.sourceforge.net/lists/listinfo/backuppc-users
Wiki:    http://backuppc.wiki.sourceforge.net
Project: http://backuppc.sourceforge.net/