Proxmox 9¶

In this article

• Deployment Features
• Proxmox 9. Installation
• First Login and Basic Check
• Network: Bridge vmbr0
• Disks and Storage
• Loading ISO Images
• Create First VM (Ubuntu 25.10)
• Windows Installation
• LXC Containers: Quick Start
• Typical VM Profiles
• Connecting VMs and LXC in One Network
• Backups and Templates
• Common Problems and Solutions
• Diagnostics: Cheat Sheet
• Mini-FAQ
• Readiness Checklist for System
• Ordering a Server with Proxmox 9 using API

Deployment Features¶

Proxmox 9. Installation¶

After installing the server, within 15-20 minutes, the installation of the Proxmox VE service is performed. An email will be sent to the mailbox linked to the account, notifying you about the installed server and providing a link in the format https://proxmox<ID_server>.hostkey.in, which you must access to enter the web management interface for Proxmox VE:

• Login - root;
• Password - system password.

First Login and Basic Check¶

1. Open your browser → https://<server_IP>:8006 and enter the credentials:

   

2. Navigate to: Datacenter → Node → Summary - check CPU, RAM, disks, uptime.

   

3. Disable the enterprise repository if there is no subscription:

Node → Repositories → pve-enterprise → Disable. Keep pve-no-subscription:

Terminal Commands:

sed -i 's/^deb/#deb/g' /etc/apt/sources.list.d/pve-enterprise.list || true

apt update

Network: Bridge vmbr0¶

The bridge vmbr0 is a virtual "switch" to which VMs are connected. It is bound to a physical interface (e.g., ens18/eno1).

Through the Web Interface¶

1. Node → System → Network.

   

2. Verify that vmbr0 exists. If it does not exist or is not configured - Create → Linux Bridge:

   • Name: vmbr0
   • IPv4/CIDR: specify your static IP in the format X.X.X.X/YY (leave blank if using DHCP);
   • Gateway (IPv4): default gateway (usually X.X.X.1) (do not enter if using DHCP);
   • Bridge ports: your physical interface, for example ens1;
   • Save → Apply configuration:

   

Through CLI (if web access is lost)¶

Example /etc/network/interfaces (ifupdown2):

auto lo
iface lo inet loopback
auto ens18
iface ens18 inet manual
auto vmbr0
iface vmbr0 inet static
    address 192.0.2.10/24
    gateway 192.0.2.1
    bridge-ports ens18
    bridge-stp off
    bridge-fd 0

ifreload -a

Common Errors:

• Incorrectly specified bridge-ports (wrong physical interface) → network "disappears". Correct the interface and execute ifreload -a.
• Wrong gateway or subnet entered → local connection exists but no internet access.

Disks and Storage¶

Add a Second Disk for VM Storage¶

1. Node → Disks: ensure that the new disk is visible (e.g., sdb).

   

2. Option A - LVM-Thin (convenient for snapshots):

   • Disks → LVM-Thin → Create: select the disk → specify VG name (e.g., pve2) and thin-pool (e.g., data2).

   

   • The storage will appear in Datacenter → Storage.

   

3. Option B - Directory:

   • Create a file system (Disks → ZFS or manually mkfs.ext4), mount to /mnt/...

   

   • Datacenter → Storage → Add → Directory → path /mnt/... → enable Disk image, ISO image (as needed).

Loading ISO Images¶

ISO images can be loaded in two ways.

A. Through the Web Interface¶

1. Datacenter → Storage → (select storage with type ISO, e.g., local) → Content.
2. Upload → select local ubuntu-25.10-live-server-amd64.iso → wait for upload completion.

B. Through the Node (CLI)¶

Example of downloading Ubuntu 25.10 ISO to local storage:

cd /var/lib/vz/template/iso

wget https://releases.ubuntu.com/25.10/ubuntu-25.10-live-server-amd64.iso

Create First VM (Ubuntu 25.10)¶

Example: Ubuntu Server 25.10 (VPS with 2 vCPU)

Click Create VM (top right):

General: Leave ID as default, Name - ubuntu2510 (or your own):

OS: select ISO ubuntu-25.10-live-server-amd64.iso, Type: Linux:

System:

• Graphics card: Default;
• BIOS: OVMF (UEFI);
• Machine: q35;
• SCSI Controller: VirtIO SCSI single;
• (Optionally) enable Qemu Agent in Options after VM creation (see below):

Disks:

• Bus/Device: SCSI;
• SCSI Controller: VirtIO SCSI single;
• Storage: your LVM-Thin/Directory;
• Size: 20–40 GB (minimum 10-15 GB for testing);
• Discard (TRIM): enable on a thin-pool:

CPU:

• Sockets: 1;
• Cores: 2 (according to your VPS);
• Type: host (best performance):

Memory:

• 2048–4096 MB. You can enable Ballooning (e.g., Min 1024, Max 4096):

Network:

• Model: VirtIO (paravirtualized);
• Bridge: vmbr0;
• If VLAN is needed: VLAN Tag:

Confirm: check settings, mark Start after created and click Finish:

OS Installation:¶

1. Start the VM → Console (noVNC) → Try or Install Ubuntu:

   

2. Installer:

   • DHCP/static IP as needed;
   • Disk: Use entire disk;
   • Profile: user/password;
   • OpenSSH server: enable.

3. Reboot and log in via console/SSH.

Post-install:¶

sudo apt update && sudo apt -y full-upgrade

sudo apt -y install qemu-guest-agent

sudo systemctl enable --now qemu-guest-agent

Boot Order: if booting from ISO - Options → Boot Order → move scsi0 above cdrom.

Windows Installation (for More Powerful Nodes)¶

Suitable for nodes with ≥4 vCPU/8 GB RAM. On weak VPS, Windows may work unstably.

1. ISO: Download the ISOs of Windows Server (2019/2022/2025) and virtio-win.iso (drivers) in Storage → Content:

   

2. Create VM → OS: Microsoft Windows, select the installation ISO image. The option Add additional drive for VirtIO drivers allows you to add a second CD with drivers:

   

3. System: BIOS OVMF (UEFI);

   • Machine: q35;
   • If necessary, enable Add EFI Disk and Add TPM (for new Windows versions). If it doesn't start - try SeaBIOS and disable EFI/TPM:

   

4. Disks:

   • Bus: SCSI;
   • Controller: VirtIO SCSI;
   • Size: 40–80 GB;
   • Enable IO Threads:

   

5. CPU: 2–4 vCPUs;

   • Type: host:

   

6. Memory: 4–8 GB:

   

7. Network: Model VirtIO (paravirtualized), Bridge vmbr0:

   

8. Confirm: Complete VM creation by clicking Create, then in Hardware → CD/DVD Drive, attach a second ISO - virtio-win.iso:

   

9. Windows Installer: On the disk selection step, click Load Driver → specify the CD with VirtIO (vioscsi/viostor). After installation - set network drivers in Device Manager (NetKVM):

   

10. Guest Agent (optional): Install Qemu Guest Agent for Windows using virtio-win ISO:

    

Troubleshooting Windows:¶

• Black screen/doesn't boot: Change OVMF → SeaBIOS, disable EFI/TPM.
• No network: Ensure NIC = VirtIO and NetKVM driver is installed.
• Disk slowdowns: Make sure the disk = SCSI + virtio driver.

LXC Containers: Quick Start¶

Pre-made templates with minimal software are available in the template storage.

1. Datacenter → Storage → (select storage with type Templates)→ **Content** → **Templates**. Download, for example:ubuntu-25.04-standard_*.tar.zst` or another needed template:

   

2. Click Create CT:

   • General: Specify ID/Name, Unprivileged container = Enabled (safer by default). Set password root or SSH key.

   

   • Template: Select the downloaded template.

   

   • Disks: Storage/Size (for example, 8–20 GB).

   

   • CPU/RAM: According to the task (for example, 1 vCPU, 1–2 GB RAM).

   

   • Network: Bridge vmbr0, IPv4 = DHCP (or Static if needed). VLAN Tag as necessary.

   

Tip for Network: If you are using NAT on vmbr1, then set it and specify the static IP.

- **DNS**: Default from the host or your own.

- **Features**: Optionally enable `nesting`, `fuse`, `keyctl` (depends on applications in the container).

• Start at boot/Start after created: as desired.
• After starting: log in via SSH and install software from the template or packages:

  apt update && apt -y upgrade
  

  

In LXC, Qemu Guest Agent is not needed. Mounting host directories is done through MP (Mount points).

Typical VM Profiles¶

• Ubuntu/Debian (Web/DB/Utility): SCSI + VirtIO, UEFI (OVMF), 1–2 vCPUs, 2–4 GB RAM, disk 20–60 GB; enable Qemu Guest Agent.
• Lightweight Services (DNS/DHCP/Proxy): 1 vCPU, 1–2 GB RAM, disk 8–20 GB.
• Container Hosts (Docker/Podman): 2–4 vCPUs, 4–8 GB RAM; separate disk/pool for data.

Alternative to ISO: You can use Ubuntu 25.10 Cloud-Init images for quick cloning with auto-configured network/SSH. Suitable if you plan to have many similar VMs.

Connecting VMs and LXC in One Network¶

Basic Variant (One Subnet):¶

1. Ensure all VMs/containers have Bridge = vmbr0 (or vmbr1).
2. If using a DHCP network - addresses are assigned automatically, if static - specify IPs in one subnet (for example, 10.10.0.2/24, 10.10.0.3/24) and common gateway 10.10.0.1.
3. Optionally. VLAN: Specify VLAN Tag in the network card settings of VMs/CT and ensure that the switch's uplink allows this VLAN.
4. Inside the OS, check that the local firewall does not block ICMP/SSH/HTTP.
5. Test: From Ubuntu VM ping <IP-LXC> and vice versa; ip route, traceroute will help with issues.

When Different Subnets:¶

• Proxmox itself does not route between bridges. A router (separate VM with Linux/pfSense) or NAT on the host is needed.
• Simple NAT on the host (example):

Enable forwarding:

sysctl -w net.ipv4.ip_forward=1

iptables -t nat -A POSTROUTING -o vmbr0 -j MASQUERADE

Backups and Templates¶

• Backup: Datacenter → Backup or Node → Backup - set up vzdump schedule (storage, time, snapshot/stop mode):

  

• VM Template: After basic VM setup → Convert to Template. Creating new VMs through Clone saves time and eliminates errors:

Common Problems and Solutions¶

"Web Interface Disappeared" (GUI Not Opening)¶

Check if the node is accessible via SSH. On the node, execute:

systemctl status pveproxy pvedaemon pve-cluster

journalctl -u pveproxy --no-pager -n 100

systemctl restart pveproxy pvedaemon

Lost Network After Bridge Editing¶

Connect via console (through provider/VNC/IPMI). Check /etc/network/interfaces and apply:

ifreload -a

ip a

ip r

VM Does Not Connect to Internet¶

• Ensure inside the VM correct IPs/mask/gateway/DNS are specified.
• Check that the Bridge of the VM's network adapter is - vmbr0 (or nat/vmbr1).
• If VLAN is used - specify VLAN Tag in NIC VM settings (Hardware → Network Device → VLAN Tag), and on the switch uplink allow this VLAN.

ISO Does Not Boot / Installer Not Visible¶

• Check Boot Order (Options → Boot Order) and that the correct ISO is connected.
• For UEFI, check if Secure Boot in the guest OS is not enabled if the ISO does not support it.

High Load/Disk "Clog"¶

• Use VirtIO SCSI and enable IO Threads for intensive disk use.
• Do not store backups on the same thin-pool that holds operational disks - better have a separate storage.

"Disconnected" Webcam/USB Device in VM¶

• For USB passthrough, use Hardware → USB Device. If the device stops responding - Stop/Start VM or reconnect USB on host. Sometimes disabling Use USB3 helps with compatibility.

Updates and Reboot¶

apt update && apt full-upgrade -y

reboot

Diagnostics: Cheat Sheet¶

Node network:

ip a; ip r; ping -c3 1.1.1.1; ping -c3 google.com

systemctl status pveproxy pvedaemon pvestatd pve-cluster

journalctl -u pveproxy -n 200 --no-pager

df -h | sort -k5 -h
lvs; vgs; pvs

cat /etc/pve/storage.cfg

qm list; qm config <VMID>; qm status <VMID>

qm stop <VMID> --skiplock; sleep 2; qm start <VMID>

Mini-FAQ¶

Q: Can vmbr0 be renamed? A: Yes, but it's not recommended on a production node - it's easier to leave vmbr0 and add additional bridges (vmbr1) as needed.

Q: Where do ISOs lie by default? A: In the local storage: /var/lib/vz/template/iso.

Q: What distinguishes local from local-lvm? A: Local - a regular directory for ISO, container templates, etc. local-lvm - LVM-Thin for VM/container disks with snapshots.

Q: How to quickly clone a VM? A: Turn an exemplary VM into a Template, then Clone → Full/Linked.

Q: How to safely scale CPU/RAM of a VM? A: Shut down the VM and change resources; for Linux part parameters can be changed "on the fly", but better planned.

Readiness Checklist for System¶

• Access to https://<server_IP>:8006 exists;
• vmbr0 is set up and internet from node;
• ISOs are loaded into storage;
• First VM is created and installed;
• Qemu Guest Agent is enabled;
• Backup is configured (vzdump schedule);
• Updates are checked.

Ordering a Server with Proxmox 9 using API¶

To install this software using the API, follow these instructions.

Some of the content on this page was created or translated using AI.