Virtualbox virtual machines keep aborting

[sjb]

We use Virtualbox to run our admin and curriculum servers on a Windows 7 machine. The host machine has 16gb RAM and is powerful enough to do the job. Everything was working fine until I had a bit of a disaster a few weeks ago and accidentally deleted the directory containing the virtual machines (stupid, I know!). Window recovered the curriculum server from an automatic backup but overwrote the admin server with a new backup whilst it was in the process of being recovered. I set up a new virtual machine, reinstalled server and reovered the admin files from a remote backup. However, ever since this was completed the virtual machines randomly abort on a fairly regular basis (once or twice a day, either one or both machines). We didn't have this problem before, so I am wondering if it is something to do with the setup of Virtualbox - perhaps the allocation of RAM between the server host and the two virtual machines (both running Server2012). Could anyone advise me what the optimum allocation would be please? Or is there anything else obvious that might be causing the machines to abort?

 

Thanks

 

Sue

[3s-gtech]

Erk.

 

I would advise moving to another hypervisor if these are production servers. This sounds like a recipe for a serious amount of work and pain.

 

What do you mean by admin and curriculum servers? Are these DCs? File hosts? Server-room-in-a-box setups?

[sjb]

They are DCs - one runs active directory and curriculum files, the other hosts SIMS and our admin files. We have been using Virtualbox for over a year and never had an issue before.

[3s-gtech]

Virtualbox is great, no doubt, and it is capable of being a great hypervisor due to its scripting abilities. The combination of desktop hardware running it and mission critical services hosted on there would just make me very nervous.

 

See if you can clone the VMs to another virtualbox host of possible. That will help you to eliminate hardware and host issues. You can also use VMWare Converter to pull them onto an ESXi host which may be more resilient (the host OS is integral to the hypervisor, rather than an application on a host OS).