Virtual Machines Backups

[adam9c1]

Not sure if this should be in the Networking or in another forum.

How do you handle vMware Fusion or Parallels virtual machine backups?
I know it is not recommended to have them backed up via TM as the image files constantly change.

 

[chrfr]

Not sure if this should be in the Networking or in another forum.

How do you handle vMware Fusion or Parallels virtual machine backups?
I know it is not recommended to have them backed up via TM as the image files constantly change.

Backing them up with any process is going to be problematic if the virtual machine is running when the backup runs. The best process would likely be using some native backup system inside the VM itself.
Time Machine running on the host will work fine as long as the VM is shut down when the backup runs.

 

[EssentialGadget]

Not sure if this should be in the Networking or in another forum.

How do you handle vMware Fusion or Parallels virtual machine backups?
I know it is not recommended to have them backed up via TM as the image files constantly change.

I run my virtual machines (VMs) on a rMBP that I backup to time machine once every two weeks or more often if necessary. My VMs are backed up by TM at this time. My 2 VMs are around 30GB each. I wouldn't include them in the TM if it is going to run hourly.

In addition, if I am making extensive changes to the VM, then I will make a local copy as a backup. I have a 1TB SSD in this unit.

Monthly, I copy the VMs to a Mac Mini with RAID and TM RAID that I use as a file server. On this server they are also uploaded to DROPBOX.

All my work files including those generated by the virtual machines are written to dropbox directories. For example, my Windows Quicken backs up to a Dropbox directory. Dropbox replicates them to two iMacs and the Mac Mini file server. All of which TM those files. This is why I don't feel the need to continuously TM my rMBP.

I am considering adding CrashPlan. Also considering adding some Superduper or Carbon Copy Cloner strategies.

Hope this helps.

 

[blacka4]

for my Vm backups, I have the VM shut down nightly at midnight, least possible user interaction, the host does the backup and spins up the VM afterwards.

it works for me right now, but I am looking at possibly doing a hypervisor host backup sometime to back up multiple machines

 

[jasnw]

I exclude the VM files (in ~/Documents/Virtual Machines.localized for VMWare) from TM and back them up with the rest of my main user directory nightly using CCC to a sparsebundle on another machine.

 

[ItWasNotMe]

Not sure if this should be in the Networking or in another forum.

How do you handle vMware Fusion or Parallels virtual machine backups?
I know it is not recommended to have them backed up via TM as the image files constantly change.

I use a two fold approach:

1. Retrospect with the server running on OSX and the (Windows etc.) clients running inside each VM to do daily backups of selected files inside the VM. (It will backup open files but the VM has to be running for the backup to run)
2. Use tmutil to change the properties of the VM so if copied elsewhere then Time Machine will back it up and then weekly do a copy of the VM to a different folder and have Time Machine back that up.

If a file is 'lost' it can be restored from Retrospect backup. If the VM is badly damaged then it can be restored from TM and the critical files then recovered to a later time from Retrospect

 

[burne]

How do you handle vMware Fusion or Parallels virtual machine backups?
I know it is not recommended to have them backed up via TM as the image files constantly change.

I'm not using Parallels but I have about a thousand VMWare, Xen and KVM VM's, almost all of them managed customer VMs.

I make backups from inside the VM, since that is the best way to make a consistent backup. For every OS and application there's a ready-made backup solution and combining them in a smart way is the way to good backups.

Many of my VMs are LAMP or WIMP. A mysql-dump, made at 04:00am, to a local disk takes care of mysql data in a consistent and restorable way. Just to be sure I keep a week of old backups. Similarly other applications that aren't consistently file based write a backup to /home/backup. Think things like PostGres, DB2, Mongo, Cassandra etc.

Next I use a normal commercial or opensource backup solution to make a backup from inside the VM. These backups follow the familiar scheme of 7 daily backups, 4 weekly backups and for most customers 3 monthly backups.

The disaster recovery procedure consists of provisioning a new vm with the same parameters as the original. After that a script automatically installs the backup software. Next I restore the latest backup or a specific one per customer request. If needed next I restore the database from the restored backup. And that results in a running copy from the night before.