Page 7 of 8

Re: Virtualization inside a virtualizer, discussions

Posted: 14. Feb 2015, 04:16
by Legorol
loukingjr wrote:@ Legorol, Is VMWare Workstation free now? (rhetorical).
I think that is a tongue-in-cheek reply that has nothing to do with my request. Regardless of price, both VMware Workstation and VirtualBox have their pros and cons. There is nothing wrong with entertaining the idea of a new feature in VirtualBox and having a polite conversation about it. I certainly didn't intend my post to be demanding anything, I was merely voicing why I would find such a feature useful.
mpack wrote:
Legorol wrote:virtualized VT-x
Do I gather that you mean a fully synthetic VT-x that works on CPUs that don't have it? I can't speak for the devs but... that ain't gonna happen. The whole point of VT-x is to do things that you can't do efficiently in software alone. It wouldn't need to exist otherwise. That and CPU's without that feature form a rapidly shrinking market of penniless users... hence ISTM targetting them is not a sensible use of resources.

OTOH, if you want to do that work yourself and then donate the code then that would certainly turn my calculation upside down.
I am not sure what the correct technical description is or how it is exactly achieved in VMware Workstation. However, I wasn't referring to physical CPUs that don't have VT-x at all. I was trying to refer to a guest seeing a virtual CPU that has VT-x capabilities, which I think is what's needed to run a hypervisor in a VM.

VMware Workstation has an option that says "Virtualize Intel VT-x/EPT or AMD-V/RVI". It still requires the physical (host) CPU to have VT-x. Once you enable this however, the guest in the VM then sees a (virtual) CPU that apprently has VT-x, and you can run a hypervisor in the guest. I have no idea how this is achieved, whether it's some sort of pass-through of the physical CPU's VT-x feature or a synthetic VT-x provided for the guest.

When I was referring to pre-Haswell CPUs, I was referring to the "VMCS Shadowing" technology that's in the new Haswell CPUs. My limited understanding is that this technology is supposed to somehow make exposing VT-x to a guest easier. However, VMware Workstation's example shows that this is not a strict requirement for having virtualized VT-x in a guest, presumably as long as the physical (host) CPU has VT-x.

I understand that development effort for VirtualBox is limited, and I don't mean anything in this post to be construed as any kind of expectation. I would just like to add my voice to the growing list in this thread :-)

Re: Virtualization inside a virtualizer, discussions

Posted: 14. Feb 2015, 04:35
by loukingjr
It was a bit tongue and cheek but, my main point was it takes time and effort to add such a feature and is partly why VMWare and Parallels charges so they can pay developers to work on the code.

I happen to own VMWare Fusion for my Macs and it too can pass VT-x once. That is a recent feature and although I can run VirtualBox running a VM inside a Fusion guest, it's a tad slow for me. That's why I prefer running two or more guests at a time with one hypervisor rather than nesting them. But that's me. :)

Re: Virtualization inside a virtualizer, discussions

Posted: 14. Feb 2015, 14:37
by Legorol
loukingjr wrote:I happen to own VMWare Fusion for my Macs and it too can pass VT-x once. That is a recent feature and although I can run VirtualBox running a VM inside a Fusion guest, it's a tad slow for me. That's why I prefer running two or more guests at a time with one hypervisor rather than nesting them. But that's me. :)
It depends on the use-case. Some things just can't be done without nested virtualization. As I wrote in my original post, I sometimes test virtualization scenarios which require a hypervisor to be able to run in a VM. A typical scenario would be having a Windows 7 host running Windows Server 2012 in a VM, enabling Hyper-V in Server 2012 and running another VM inside that. I get acceptable performance in such a scenario. For example, I had a case the other day that required two Server 2012 VMs with 2-3 guests running in each, for a total of 5-6 nested VMs. I was able to work fine with that.

The bottom line is, it's not a question of personal preference, nested virtualization is simply a useful feature for some scenarios.

Re: Virtualization inside a virtualizer, discussions

Posted: 14. Feb 2015, 14:43
by loukingjr
I know. By "but that's me." I meant I have no use for nested hypervisors myself. Not a personal preference. :)

Re: Virtualization inside a virtualizer, discussions

Posted: 19. Feb 2015, 00:04
by nbacr
And what if do I need install Windows Server 2012 R2 Hyper-V into VirtualBox for testing?

I tried, when I was enabling Hyper-V on Windows Server 2012 R2, I had a message, it said that my CPU doesn´t support Huper-V. I checked the AMD-V and Intel VT feature was enabled.

What should be the way to implement it?

Re: Virtualization inside a virtualizer, discussions

Posted: 19. Feb 2015, 00:22
by Perryg
And what if do I need install Windows Server 2012 R2 Hyper-V into VirtualBox for testing?
It won't work because the VMX flag is not passed to the guest.

Re: Virtualization inside a virtualizer, discussions

Posted: 19. Feb 2015, 00:53
by michaln
Legorol wrote: I was merely voicing why I would find such a feature useful.
What you forgot to say is how much it would be worth to you in terms of dollars or time/effort :)

As far as I know, there's been zero customer interest in such feature so far, however "nice" it might be. And, surprise surprise, features that are worth nothing to customers are further down on the todo list.

Re: Virtualization inside a virtualizer, discussions

Posted: 19. Feb 2015, 04:39
by Legorol
michaln wrote:What you forgot to say is how much it would be worth to you in terms of dollars or time/effort :)

As far as I know, there's been zero customer interest in such feature so far, however "nice" it might be. And, surprise surprise, features that are worth nothing to customers are further down on the todo list.
I understand that the priority is on paying customers, as it should be. We are just having a friendly discussion about a feature suggestion in a thread dedicated to it, so I think there is no need to kill the discussion with the money argument.

Personally, I would love to be able to buy VirtualBox. In my opinion it is the superior product compared to VMware Workstation and would be willing to pay for it. I had a look at what the commercial options are and unfortunately Oracle doesn't seem to offer a commercial option for individuals. I found the option to buy Enterprise licenses (minimum of 100 users) and Premier Support contract (designed for organisations), but nothing else.

It is frustrating that there seems to be no option for an individual to purchase, yet when I try to voice any suggestions or concerns, I am effectively told to "pay up, or else...".

Have I missed a purchase option for an individual? If so, please direct me to it.

Re: Virtualization inside a virtualizer, discussions

Posted: 19. Feb 2015, 04:58
by loukingjr
Legorol wrote: I understand that the priority is on paying customers, as it should be.
"paying customers" is redundant.
We are just having a friendly discussion about a feature suggestion in a thread dedicated to it, so I think there is no need to kill the discussion with the money argument.
Actually there is nothing to discuss. Some users would like nested hypervisors. Discussion over. :)
Personally, I would love to be able to buy VirtualBox.
Why? It won't get you nested hypervisors any quicker. What you need to do is find 10,000 others who want the feature to buy VirtualBox.

I don't want to seem mean but you are taking the phrase "beating a dead horse" to a new level.

Re: Virtualization inside a virtualizer, discussions

Posted: 19. Feb 2015, 14:35
by michaln
Legorol wrote:It is frustrating that there seems to be no option for an individual to purchase, yet when I try to voice any suggestions or concerns, I am effectively told to "pay up, or else...".
No, you're told "it's free, take it or leave it" :)
Have I missed a purchase option for an individual? If so, please direct me to it.
No. The current pricing is in this document http://www.oracle.com/us/corporate/pric ... 070592.pdf The minimum is 50 licenses (and it costs a bit more to get support).

The reality that Oracle is not equipped to deal with individual users, and that for a product such as VirtualBox with a very wide range of uses, it's not feasible to provide official support for every single user. For an individual, I'm sure it would be great to get every obscure bit of hardware or software supported, but it would mean effectively halting the development of any new feature because we definitely wouldn't have time for anything else.

And when it comes to deciding what new feature does or doesn't get implemented, obviously customers get to influence that, and features that benefit a majority of users get prioritized. If and when nested virtualization will make it to the top of the todo list is not something we can comment about.

Re: Virtualization inside a virtualizer, discussions

Posted: 27. Mar 2015, 15:34
by pncampbell
I've got virtual box 4 running inside vmware player 7 i.e. ->
Host is windows 7 x64 on i7-4790k, vmware guest os is ubuntu with virtualbox installed, guest os on virtualbox is some mini linux web appliance (I just wanted a small vm to test this out).

Basically it "just worked", once I had ubuntu running on vmware I switched to the guest vm and installed virtualbox as if I was working in ubuntu natively. Didnt need to play with any settings anywhere.

Re: Virtualization inside a virtualizer, discussions

Posted: 17. Jun 2015, 15:06
by ohaya@yahoo.com
walkamongus wrote:I'm running a Virtualbox VM guest inside a Virtualbox VM host (via Vagrant) and the VM-inside-the-VM boots up fine. The issue I'm running into is that I am unable to access the VM guest on the host-only network from the virtualized host. The vboxnet0 interface is up in the virtualized host and it has the default gateway IP address that I can correctly ping, however I am unable to access the VM guest that is on that network.

I tried this setup by virtualizing the host on both VMware and KVM and had the exact same issues, which is why I think the issue is not running Virtualbox inside Virtualbox.

Has anyone encountered this issue before?

I'm also running VBox inside VBox. The first physical box in Win7 with VBox 4.3.20. I run a Win 2008 R2 guest under that box, and have VBox 4.3.20 on that inner guest and that inner VBox has several Win7 32 bit guests.

I can access the Win7 guests from the inner Win 2008 R2 (i.e., I can access the innermost guests from the guest that is hosting the inner VBox.

I CANNOT access the inner Win7 guests from any machine (either virtual or real) that is outside of inner Win 2008 R2.

Re: Virtualization inside a virtualizer, discussions

Posted: 21. Aug 2015, 16:33
by AnrDaemon
AnrDaemon wrote:Anyone had a success using LXC under Linux guest?

I know it's not quite the virtualization, but close enough to ask before try. :roll:
Answering my own question, the only caveat I've found is to allow all passthrough traffic on the VirtualBox network layer. Then LXC works properly in all networking modes I tested.

Re: Virtualization inside a virtualizer, discussions

Posted: 17. Sep 2015, 07:19
by adelnet10
thx

Re: Virtualization inside a virtualizer, discussions

Posted: 13. Nov 2015, 11:25
by MadNoblex
How many level are possible?