networking – How can I ping a VM on a hypervisor from outside the host?

I have a CentOS 7 hypervisor running. The hypervisor IP is 192.168.195.10, and I created a VM with the IP address of 192.168.122.136 and the gateway of 192.168.122.1.

I can ping from hypervisor to the VM.

When I try to connect other computers to the network I can ping from it to the hypervisor, but I cannot get to the VM and not to the gateway of the VM.

  • Host to VM: Good
  • Outside to Host: Good
  • Outside to VM: No Good

Can someone give me direction to solve this issue?

Hypervisor / server that I can install on laptop with three OS’es: Windows 10, Ubuntu server and Ubuntu desktop

I would like to install Hypervisor on my Gamming laptop with RTX 2060 so I can have 3 OS’es:

  1. Windows 10
  2. Ubuntu server 20.04 LTS
  3. Ubuntu Desktop 20.04 LTS

I am interested in passthrough of Nvidia RTX 2060 for guest Windows nad Ubuntu Desktop and get near native performance at guest if possible. Server should work all the time when laptop is on in background. There wont’e much load at server. Mostly I would work at Ubuntu Desktop, but some time I would like to have also Windows on. So at peak I will have 3 OS’es running simultaneously and switch between them easily.
I would like to work directly on the laptop so I can operate on the same machine with three OS’es with GUI.

Is it possible? Is it worth to do that considering pitfalls? How can I do that?

cloud computing – Bare metal as a service – hypervisor and gpu?

What services are available for customizable cloud bare metal administration? Or do I have options for buying and using my own equipment with a public IP? In particular, I’d like something on which I can run a type 1 hypervisor to virtualize several servers, some with a public IP, and at least one serving a remote desktop with a decent GPU for content creation/gaming.

I have several VPSes on DigitalOcean where I host client websites. I’ll be launching a SaaS app later this month, also on a DigitalOcean VPS. I also use PaperSpace with Parsec as a remote desktop, where I do some gaming as well as video editing and graphics-intensive development. I also take advantage of PaperSpace’s gigabit internet for network-intensive tasks, since gigabit internet comes at a premium in my area of the US.

It strikes me that I can consolidate a lot of that/all of that onto a single server running a hypervisor. That would give me more control, too (and maybe more headache?). Using my own equipment, however, poses three problems:

  • the need to pay for higher internet connection speeds
  • service outages if I lose power, need to move equipment, or pack up and move to another house
  • lack of access to a public-facing IP address. A cloud server would seem to solve these three problems.

I’m not asking anyone to Google for me. OVH Cloud, which has some great options; Cyxtera looks to be a bit pricey and emphasizes enterprise. I don’t see any options to add a dedicated GPU, though. Oracle offers servers with GPUs, but I honestly can’t tell if its something I can install a hypervisor on or not… (I’m a hobbyist, not a sysadmin–this is all pretty foreign territory!). I’ve seen other cloud services offering what I’m looking for with a NVIDIA Tesla GPU, which I understand is not super great for gaming (same for Quadro–I’m told they emphasize precision over speed, where I think I’m preferring speed here).

  • Is cloud administration the right route to go here? (I’ve honestly considered just having my own equipment on a private network and keeping my DigitalOcean VPSes separate–but I like the idea of everything being in one place).
  • Is this even reasonable/feasible?
  • I saw a Linus Tech Tips video where he did something like this, and he slotted in a few gaming-class GPUs. Of course, he wasn’t hosting publically-reachable web sites and apps. Can I get this level of configuration on a cloud-managed service?

kvm virtualization – How much disk space is required to install KVM hypervisor on a CentOS machine?

If you want to install KVM, QEMU, libvirt, virt-manager, your available space is more than enough. The absolute minimum is KVM and QEMU which take very little space.
If you want to install oVirt, a vSphere or MS SCVMM like web console, you can also proceed, but probably you’ll soon need more space – it also needs a DB backend.

But you should be careful regarding the locations of VMs – virtual disks and ISOs, that’s the ones taking space. Probably you’ll have to change the default location of the libraries.

Understanding type-1 hypervisor architecture. How are calls to hardware made?

When using a type-1 hypervisor, I want to understand few things better.

  1. Does the hypervisor give direct control of the requested hardware to the guest. Things like allocation of number of cpu cores, memory etc. Or the hypervisor sits between the guest/virtual os and hardware and takes the responsibility of translating the call to underlying hardware.

  2. How are unique resources like keyboard / display made available to the guest. Is it just contextual, based on which OS is active gets to use these resources like with Guest running on VirtualBox/VMware running on Host. (I never got to see a type-1 hypervisor in action, so this crazy question.)

How to use openstack with xenserver as hypervisor?

I have a physical server which has XenServer installed. How do I architect this so that I can use OpenStack (via DevStack).

At the moment the setup is as follows (super simplified version…);

Physical Machine
1x Physical Hard Drive – XenServer Installed (aka. DomU)

    • VM X (Contents not related to OpenStack)
    • VM Y (Contents not related to OpenStack)
    • VM Z (Ubuntu Server 18.04 with DevStack Installed and XenTools Installed)
      After installing OpenStack on VM Z int he example above, while I can login to the Horizon dashboard, I can’t spin up any Instances which makes me think that I’ve got something setup incorrectly.

I’ve also been reading conflicting/unclear information online that is also suggesting that you may or may not need to install some Xen plugin on DomU to enable some form of communication with the OpenStack Controller node. Not sure how accurate that information is that I’ve read though as my understanding is that everything is setup with DevStack on VM Z which would be classified as the Controller Node and the Compute Node and the Identity Node etc. in this setup.

Thoughts?

VLAN aware bridging for KVM hypervisor on debian 10 (buster) and later

I am setting up a new KVM hypervisor and the VMs will need to connect to several VLANS (one VLAN per VM). However I can’t find good documentation or instructions on how to configure the hypervisor or the VMs – perhaps because VLAN filtering and VLAN aware bridges are still new (from 2018 i think).

So I look for examples and best practices on setup and maintenance. Do I need special packages? Bonus points for Ansible receipts.

kvm virtualization – How do remove the default storage pool from a libvirt hypervisor, so that even after libvirtd restarts there is NO storage pool

I want to remove the default storage pool from my virt-manager AND NOT HAVE IT COME BACK BY ITSELF, EVER. I can destroy it and undefine it all I want, but when i restart libvirtd (for me thats “sudo systemctl restart libvirtd” in an arch linux terminal window), and restart virt-manager, the default storage pool is back, just like Frankenstein.

I don’t want a storage pool of any kind. I simply want to move from the dual-boot I have now (arch linux and windows) to running the two OS simultaneously. I intend to provision two physical disk partitions on the host to be disks on the guest, and I can do this via the xml that defines the domain.

Or am i required to have a storage pool no matter what?

centos7 – CPU utilization of the KVM hypervisor – CentOS 7

We have updated our CentOS 6 KVM hypervisors to CentOS 7.
The CPU utilization of the hypervisors has been much higher since the upgrade.
We migrated the VMs from the hypervisor to another, then upgraded to CentOS 7 and migrated them back.

This is a mixed hypervisor with Linux and Windows guests. We limit it with vcpu_quota.
We have no idea why the load has increased so much. See picture for the diagram of the CPU utilization.
Imgur: HV CPU load

Do you have any idea why the load has increased so much?
The same guests are there.