0

I am considering running an unraid server (https://unraid.net/) using the virtual machine manager (basically a front end for KVM) rather than many separate machines to do tasks in my network.

My concern is this, having all these VMs/Docker images on a single machine introduces a single point of failure.

My question, then, is how best to mitigate this? I would like to have a script (unraid is a custom linux system) that backs everything up to an ssh server also residing on the same network but a seperate physical machine.

I am also aware that should my office burn down, etc. I would be without backups.

I need strategies and a specific implementation for this scenario.

How do other admins running a virtual infrastructure deal with this?

Improve this question
3
  • You should take the same approach to backup your VMs as you would any file. A question seeking a script to do this would be out of scope.
    – Ramhound
    Commented Sep 29, 2018 at 10:07
  • Why a second SSH server, a NAS would do just as well. Doing backups of the real VM images is costly, a minor change requires to copy the whole image. Either do snapshots, or have the VMs run their own backup to the NAS, and have some way to reprovision them quickly (Vagrant or else).
    – xenoid
    Commented Sep 29, 2018 at 13:44
  • @xenoid Because I have a second server with sufficient space that runs an SSH/SFTP server. By "snapshots" I assume you're referring to btrfs, but the VMs use their own filesystems (eg Ubiquiti UCRM, UNMS). Being these two are linux systems it may be possible to run backup on the CLI of these systems however it needs to be done while they are running.
    – CellPointNetworks
    Commented Sep 29, 2018 at 15:48

1 Answer 1

Reset to default
0

Being that all the VMs are Linux based, I elect to use btrfs and back up using the snapshots feature of the filesystem.

This is explained in detail here.

Improve this answer

Not the answer you're looking for? Browse other questions tagged .