Guide · Virtualization
Many teams end up with more than one hypervisor: a KVM host they built, a Proxmox VE cluster they inherited, an ESXi server that runs Windows workloads. You do not have to consolidate them to manage them well. This guide covers the daily jobs, where the platforms differ, and how to hand machines to customers or colleagues from one place.
In this guide: The short answer · A worked scenario · Daily jobs · Where platforms differ · Three decisions · Questions
Nodekeeper’s virtualization management (Pro and up) connects to KVM hosts, Proxmox VE and VMware ESXi and gives you the same workflow for the jobs you do every day: install a system, assign an address, open a console, power a machine on or off, snapshot, back up and restore, and hand it to a user. It does not make the platforms identical, so know what differs before you plan around a feature.
A hosting team sells small Linux VPS plans on KVM, took over a Proxmox VE cluster with a customer’s workloads, and keeps an ESXi host for a few Windows machines. Support gets the same requests on all three: reinstall a system, open a console, take a snapshot before an upgrade, restore a machine, give a customer access to just their VM. Without a shared tool each request means a different interface, a different login and a different set of habits.
With one panel, the person on support learns one workflow, the permission model is the same everywhere, and every action lands in the same operation log.
| Job | What you do in Nodekeeper |
|---|---|
| Install a system | Pick an image or upload your own ISO, set size and network, and let the panel install it. Windows and Linux images are available on KVM; ESXi supports one-click provisioning with unattended Windows install. |
| Give it an address | Assign IPs from a pool (Proxmox VE) or with IP guard (ESXi), so two machines do not end up with the same address. |
| Open a console | Use the web console from the panel, over the panel’s authenticated HTTPS connection. No separate console client or tunnel to set up. |
| Power, snapshot, back up, restore | Start, stop and restart machines, take snapshots before risky changes, and back up and restore VMs from the same screen. |
| Recover a machine that will not boot | Boot the rescue system. It runs from memory and does not wipe the machine’s disk. |
| Hand a machine over | Assign a VM to a member account. The member sees and manages only their own machines. |
| Reinstall a whole server | Rewrite a physical server from a disk image, with no hypervisor required. This one is included in every plan, Free too. |
The panel is the same; the platforms underneath are not. Plan around these differences.
| Workflow | KVM | Proxmox VE | VMware ESXi |
|---|---|---|---|
| Install systems | Many Windows and Linux images, or a custom ISO | Reinstall for LXC containers | One-click provisioning, unattended Windows install, EFI network boot |
| IP assignment | — | IP pool | IP guard |
| Web console, snapshots, rescue system | Yes | Yes | Yes |
| Port forwarding and traffic limits | Yes | — | — |
| GPU, USB and disk passthrough | GPU, whole disk | GPU, USB, disk | GPU, disk (RDM), USB auto-connect |
| System disk on distributed storage From Business | Yes | Yes | Yes, over NVMe/TCP |
| High-availability failover From Business | Yes | Yes, for VMs | Not offered |
A dash means we do not list that workflow for that platform today; ask before you plan around it. Passthrough needs IOMMU / VT-d on the host and is assigned by administrators only.
When you sell machines rather than only run them, the work around the VM matters as much as the VM: orders that trigger a build, quotas per customer, suspend and resume when an invoice is unpaid. Business adds customer quotas, the hosting API and WHMCS integration. Hosting and VPS provisioning covers that workflow.
Nodekeeper itself can also run inside a VM on ESXi or Proxmox VE, and natively on ESXi 7, 8 and 9 as a package. That ESXi build is a dedicated port; it does not mean any Linux program runs on ESXi. If you plan to run Nodekeeper on both a host and inside its guests, ask us how licences are counted first.
No. Nodekeeper connects to KVM hosts, Proxmox VE and ESXi that you already run and manages the machines on them.
Failover restarts a machine on another host of the same platform, using shared storage. Moving a machine from one platform to another is a migration you plan separately.
If a host fails, a supported VM is started again on a healthy host with its original IP and MAC addresses, which saves the manual recovery work. It applies to KVM and Proxmox VE VMs on distributed storage, with some exclusions. See VM recovery and iSCSI failover for prerequisites.
KVM, Proxmox VE and ESXi management starts with Pro. Storing VM disks on distributed storage and automatic restart on another host are Business. The pricing page lists per-node prices and what each plan includes.
Start from the virtualization page for the full list of workflows per platform, or compare plans first.
Related: virtual machine documentation · VM recovery · distributed storage management · multi-site networking
Manage websites, office PCs and servers from one panel on your own hardware.