0

I just backed up a vm using bareos (a fork of Bacula) and it appears to have worked, but I installed the file daemon on the vm itself, and it doesn't feel right; I ended up with what appeared to be a looping directory in the /var/lib/docker/devicemapper/devicemapper/data directory of the vm (the file-daemon just kept backing that up forever until the backups that filled up the directory were larger than the drive being backed up), so I excluded it and the backup completely very quickly.

Instead, should the File Daemon be halting or pausing the Virtual Box machine, and then backing up the VM disk file?

This can be accomplished using the ClientRunBeforeJob directive of a Job, and then it can be run again using the ClientRunAfterJob directive of a Job along with the vboxmanage command.

In short, which is a more reliable way to do this for bareos backups?

leeand00
  • 4,807
  • 13
  • 64
  • 106
  • What is your strategy for restoring? Actually _execute it_ once, and then try it again via the other method you've proposed. You will then know exactly what you need to know (and as a bonus you'll learn why VirtualBox is still not taken seriously). – Michael Hampton Nov 18 '16 at 16:35
  • @MichaelHampton Well I was initially going for restoring the entire VM in case I needed to, back to the vm...but really I'm just trying to learn the different ways in which one can backup and restore machines and VMs in BareOS. So I want to know how to do it both ways. Up to this point I've only really restored files back to /tmp/bareos/ – leeand00 Nov 18 '16 at 17:44

1 Answers1

2

Well, depends on you backup/restore strategy. We have 16 active virtual machines. The initial steps on our backup strategy are:

  1. shutdown the virtual machine (ACPI)
  2. backup the directory containing the virtual machine
  3. restart the virtual machine

the steps 1 & 3 assures the health of the always unstable Windows servers.

On our case, the virtual machines do not contain a large quantity of data. The container directory are about 150 GB each.

But you could make a full backup on monday and the rest of the days do several snapshots.

In our case, we prioritize the fast recovery of the full backup done every night. We do a restore simulation every month to assure the backup/restore procedures are working fine.

Patrick Mevzek
  • 9,273
  • 7
  • 29
  • 42
m0rg4n
  • 136
  • 4