This course takes you from your first Proxmox VE installation to managing it as code. It covers choosing storage and networking, KVM VMs and LXC containers, access separation with RBAC (role-based access control), backup with Proxmox Backup Server, migration from VMware ESXi, then Terraform and Ansible. It is aimed at system administrators, DevOps engineers and SREs who must run a production hypervisor. It targets Proxmox VE 9.x, built on Debian 13 Trixie, and is meant to be followed in order: concepts first, then installation, then delegation of access, then hands-on cases and automation.
What are the prerequisites for the Proxmox VE course?
Two conditions weigh on the whole path. First, the prerequisites: without fluency in a Linux terminal and some notion of IP addressing, the networking and storage lessons become frustrating. Second, the version: the course targets Proxmox VE 9.x, and configuration paths and command outputs differ from versions 7 and 8, which are still very common in production.
The path declares the KVM/libvirt course (in French) as a prerequisite, and it is not a formality: Proxmox VE is a Debian that wraps KVM and LXC behind a web interface and its own command line. What you learned there about QEMU, cloud images and cloud-init carries over here unchanged, and it is what makes readable what the interface hides. An equivalent hands-on experience of virtualization on Linux plays the same role: what matters is what you have learned, not the route by which you learned it.
Why is Proxmox VE an open source alternative to VMware?
Since Broadcom's acquisition of VMware at the end of 2023 and the overhaul of its licensing (end of perpetual licenses, forced move to subscriptions, price increases), many organizations have been looking to leave vSphere and ESXi. Proxmox VE has become one of the most serious open source alternatives: a complete type 1 hypervisor with no license cost, which runs KVM VMs and LXC containers with clustering, high availability and built-in backup.
The switch is all the easier because Proxmox ships a native Import Wizard: it imports machines directly from ESXi, both their configuration and their disks, to the storage you choose. The VirtIO drivers of a Windows guest, however, are still yours to install. The full procedure is in the guide Migrating from VMware ESXi to Proxmox VE (in French). The course gives you the essential foundations before, during and after that migration.
How is the course organized?
The course outline, in the left column, lists the lessons and their durations: this page does not repeat them, it explains the role of each module and the reason for their order. The modules are not alike, and you will not approach all of them from the same starting level.
The first module, Understanding Proxmox VE, requires no server. Read it
even before downloading the ISO, because the decisions that are the most
expensive to fix are made during installation: the storage type of the
system disk (ext4 on LVM-thin, or ZFS) and the management bridge. Going back
on them later often means reinstalling the node.
The second module, Installing and creating your first instances, is done hands on keyboard on your own node: all the following modules rely on the node you install there. It covers two installation routes, because they are not met in the same situations: the official ISO when you control the machine's boot, and installing the packages on an existing Debian 13 when you do not, at a hosting provider or in a lab built from a cloud image. The first dwells on the filesystem choice screen, the second on the four pitfalls that a cloud image adds to the official procedure.
The tour of the web interface, served on port 8006, explains the difference
between the Datacenter view and the Node view, a distinction that comes back
in every following lesson. You then create a KVM VM, then an LXC container: only
at that point do the notions of snapshot and resources make sense, because
there are finally guests to observe.
The module Mastering identities and boundaries comes next, and its place is
no accident: a delegation only makes sense on a node that already hosts
workloads to hand over. It teaches Proxmox's authorization model, where a
subject, a role and a path together decide what is allowed. The real benefit
is elsewhere than in the theory: it keeps you from working for months as
root@pam, a habit that is hard to shake afterwards.
The Hands-on cases module gathers situations met in production: running Windows 11 (firmware and TPM included), taking over a VMware ESXi estate without rebuilding everything, backing up with PBS, and making a VM's IP address show up thanks to the QEMU agent. The lessons are independent: take them in the order that serves your need, once the first two modules are done.
The Automation (IaC) module drives Proxmox VE as code rather than with the
mouse. It relies on a cloud-init template that Terraform clones to create
each machine, and on the QEMU agent to retrieve their addresses. This is where
the choices made at the start pay off: the Terraform provider works with an
API token and a least-privilege role, not as root@pam.
What will you be able to do at the end of each stage?
Here, for each module, is a self-assessment criterion that you must be able to state without lying. If, at the end of the first module, you cannot say why you would choose ZFS over LVM-thin, go back over the storage lesson before installing: an unmet objective is paid for in the next module, when you create your first VM.
- After the concepts: "I know what a node, a storage and a bridge are, and I choose my storage and my network before installing".
- After the installation: "I can install Proxmox from the ISO or on an existing Debian, navigate the interface, and create a KVM VM and an LXC container".
- After the identities: "I can delegate the operation of a pool without handing over the keys to the node, and check that a denial lands exactly where I expect it".
- After the hands-on cases: "I can set up a Windows 11 VM, import a VM from ESXi, back up with PBS and have the guest agent report IP addresses".
- After the automation: "I can clone a cloud-init template with Terraform and drive my estate with Ansible".
How does the course handle failures?
The course follows an incident-driven approach: the lessons draw on real failures, so you learn to recognize the symptoms, identify the cause, then apply the fix, exactly as in production. The table below lists the cases covered, with the lesson that handles each one.
| Symptom | Actual cause | Where it is covered |
|---|---|---|
| VM does not boot | Wrong disk bus, inconsistent EFI or BIOS firmware | Create a KVM VM |
| VM network down | Wrong bridge, VLAN or tag | Networking concepts |
| Very slow I/O | Unsuitable storage backend, saturated storage | Storage concepts |
| Snapshot that "breaks" performance | Confusion between snapshot and backup | What is Proxmox VE? |
| Unusable backup | Never tested, forgotten dependencies | Proxmox Backup Server (in French) |
| Access leak | Overly broad RBAC, shared tokens | RBAC concepts (in French) |
| "Guest Agent not running", missing IP | QEMU agent absent from the guest | qemu-guest-agent |
Where should you start, depending on your level?
Choose the entry point based on what you can do without documentation: if you have to look up how to create a Linux bridge, take the first entry, even if you have already installed Proxmox once. The two other entries do not skip lessons for convenience, they assume you already know the node, storage and bridge vocabulary used in the following guides.
- New to Proxmox: follow the lessons in order, practicing each concept before moving on to the next, starting with What is Proxmox VE?.
- Already some basics: if you can install Proxmox and create VMs, go to the storage lesson (Storage concepts), then to the hands-on cases.
- Team lead: prioritize RBAC (RBAC concepts, in French), backup with PBS and the Terraform provider, whose least-privilege role drives architecture decisions.
What does the Proxmox VE course not cover yet?
Two workstreams are planned after the path, but they are not yet published. The first, industrialization, will cover repeatable cloud-init templates, advanced ZFS, shared storage, advanced networking (bonds, VLAN trunks), the 3-2-1 backup strategy with PBS, disaster recovery planning, the Proxmox firewall, observability and performance. The second, the platform, will cover the cluster and high availability, Ceph, SDN, hardening of the host and workloads, and automation at scale. Cluster, Ceph and version upgrade incidents will arrive with it.
Which official resources should you consult?
Four addresses are enough during the course. The Proxmox wiki documentation is
authoritative on the qm and pct commands, the forum is where production
incidents are discussed, and the vendor's GitHub lets you read the tools' code
when the documentation is not enough.
- Official site: proxmox.com/proxmox-ve
- Documentation: pve.proxmox.com/wiki
- Forum: forum.proxmox.com
- GitHub: github.com/proxmox