Skip to content

Operations VM Setup ​

The operations VM is a general-purpose Docker host for monitoring, network tools, and infrastructure inventory. It runs as a guest on the physical Proxmox host, separate from the media server.

For the lab's automated provisioning, SSH account, and AI operating procedures, see Operations VM · Internal Operations.

Configuration ​

SettingThis homelabFor another homelab
GuestProxmox VM 100An unused guest ID
HostnameoperationsYour chosen hostname
OSDebian 13 (Trixie)A supported Debian release
CPU2 vCPUSize for your workloads
RAM8192 MB, ballooning disabledReserve enough for your containers
Disk64 GB on local storage flashLocal Linux storage with room for state, images, and logs
Address10.120.40.100/24A stable address on your server network
Gateway10.120.40.1Your router
DNS10.120.40.103A reachable resolver
Search domainhome.rwk.meYour internal domain, if used
Network bridgevmbr0A bridge connected to the intended network
StartupStart at boot enabledMatch your dependency and recovery plan

The lab's automated path uses a verified Debian cloud image. The manual path below uses the Debian installer and produces the same service-host layout without requiring access to private scripts. Replace lab-specific addresses, storage IDs, and network choices consistently.

Create the guest ​

Before creation, check Proxmox inventory, DHCP reservations, and network assignments to confirm the guest ID and address are free. An address failing to respond to ping does not prove it is unused.

  1. Download a Debian 13 amd64 installer ISO from Debian and verify its published checksum. Upload it to Proxmox storage supporting ISO images.
  2. Select Create VM in Proxmox. Choose the guest ID and name from the table, then select the installer ISO.
  3. Use the q35 machine type, a VirtIO SCSI single controller, and enable the QEMU Guest Agent option. The lab's automated configuration uses the x86-64-v2-AES CPU model, one socket, and two cores; ensure the model is supported by your physical CPU.
  4. Create a 64 GB SCSI disk on flash or your chosen local guest storage. The lab enables discard, SSD emulation, and IO thread for this disk. Do not initialize an existing guest's disk to follow these instructions.
  5. Allocate 8192 MB RAM with ballooning disabled for the lab configuration.
  6. Add a VirtIO network adapter on the intended bridge. Match its VLAN settings to the physical host and switch configuration. Do not add a VLAN tag blindly if the bridge already presents the intended untagged network. If enabling Proxmox's guest firewall, allow the required management and service traffic.
  7. Boot the installer and open the guest console.

Install Debian ​

Choose the hostname and timezone, create a normal administrator account, and install a minimal system with SSH server and standard system utilities. A desktop environment is not required. Store account credentials in your password manager and ensure the account has sudo access before relying on it for administration.

Configure the address, gateway, and DNS from the table using Debian's installer or the installed network manager. Reserve or exclude the address in DHCP. Keep the Proxmox guest console available when changing network settings.

After installation, remove the installer ISO from the boot path and boot from the virtual disk. Confirm the guest's actual address, route, and DNS behavior:

sh
hostnamectl
ip -brief address
ip route
getent hosts download.docker.com

The lab uses Pi-hole at 10.120.40.103 for DNS. If it is unavailable, package downloads and hostname-based checks can fail even when the operations VM is running. Restore that dependency or deliberately choose a reachable bootstrap resolver during provisioning.

Prepare the operating system ​

Run inside the guest from a sudo-capable administrator account:

sh
sudo apt update
sudo apt full-upgrade
sudo apt install ca-certificates curl dnsutils gnupg jq locales openssh-server qemu-guest-agent sudo
sudo systemctl enable --now ssh
sudo systemctl start qemu-guest-agent

Confirm the QEMU Guest Agent option is enabled in Proxmox as well as installed in Debian. If its virtual device is missing after enabling the option, perform a controlled guest shutdown and start. Verify the service is active and Proxmox can query the guest. Review Proxmox guest-agent configuration for the host-side options.

sh
systemctl is-active qemu-guest-agent

Set your desired timezone with sudo timedatectl set-timezone America/Chicago, substituting your region as needed. Containers also receive their own TZ setting in each project's environment file.

Configure key-only SSH ​

Install your selected administrator public key while preserving any existing authorized keys. Verify a separate key-authenticated connection before disabling password login. Keep the Proxmox guest console and the original session available as recovery paths.

For a fresh guest with a working non-root sudo account, create /etc/ssh/sshd_config.d/00-key-only.conf containing:

text
PubkeyAuthentication yes
PasswordAuthentication no
KbdInteractiveAuthentication no
PermitRootLogin no

Validate with sudo sshd -t and inspect the effective configuration with sudo sshd -T. Include order and Match rules can affect which values apply. Reload using sudo systemctl reload ssh, then prove a new key login succeeds and password-capable authentication is rejected. The lab's exact account and reviewed hardening assets are documented internally.

Install Docker Engine and Compose ​

Use Docker's Debian repository rather than the Ubuntu repository. On a fresh guest, add the signing key and repository:

sh
sudo install -m 0755 -d /etc/apt/keyrings
sudo curl -fsSL https://download.docker.com/linux/debian/gpg -o /etc/apt/keyrings/docker.asc
sudo chmod a+r /etc/apt/keyrings/docker.asc
sudo tee /etc/apt/sources.list.d/docker.sources >/dev/null <<EOF
Types: deb
URIs: https://download.docker.com/linux/debian
Suites: $(. /etc/os-release && echo "$VERSION_CODENAME")
Components: stable
Architectures: $(dpkg --print-architecture)
Signed-By: /etc/apt/keyrings/docker.asc
EOF
sudo apt update
sudo apt install docker-ce docker-ce-cli containerd.io docker-buildx-plugin docker-compose-plugin
sudo systemctl enable --now docker
sudo docker run --rm hello-world
sudo docker compose version

If Docker already exists, inspect the packages and repository configuration before changing it. See Docker's Debian installation guide for conflicting-package removal and supported releases.

Use sudo for Docker commands, or add your chosen deployment account to the docker group and start a fresh login. Membership grants root-equivalent host control. The service guides assume the account can run docker compose without sudo.

Limit Docker logs ​

Create or merge these settings into /etc/docker/daemon.json, preserving unrelated existing options:

json
{
  "log-driver": "json-file",
  "log-opts": {
    "compress": "true",
    "max-file": "5",
    "max-size": "10m"
  }
}
sh
sudo dockerd --validate --config-file=/etc/docker/daemon.json
sudo systemctl restart docker

Restarting Docker interrupts workloads already running on the guest. Configure logging before the initial service deployment. Existing containers need recreation to inherit changed logging defaults.

Prepare service storage ​

Keep application databases and container state on the guest's local Linux filesystem. The operations services do not require a NAS mount. VM backups can still use NAS storage managed by Proxmox.

Create each project directory when deploying that service, and give the deployment account ownership. For example, run as that account:

sh
sudo install -d -o "$(id -u)" -g "$(id -g)" -m 0750 /docker/uptime-kuma

Each application gets its own Compose project under /docker. Preserve existing environment files and data during updates. Some application images require specific data-directory owners; follow the individual service's guide rather than applying recursive ownership changes to all of /docker.

Deploy the operations services ​

ServicePurpose
IT-ToolsBrowser utilities
LibreSpeedLocal client throughput tests
RackPadInfrastructure inventory
Speedtest TrackerScheduled internet performance history
UniFi ToolkitUniFi administration tools
Uptime KumaAvailability monitoring

Start with Uptime Kuma if you want monitoring ready while adding the other services. Each guide owns its application configuration. Configure DNS and Nginx Proxy Manager separately for each HTTPS endpoint and allow the proxy to reach the corresponding backend port.

The private homelab repository retains the creation script, provisioner, Compose files, and environment templates. Its automated host provisioner prepares Uptime Kuma's files but does not start the application; complete the service guide afterward. This lab also uses a dedicated operations deployment runner, configured through the internal runbook.

Backups and reboot checks ​

Configure an off-host backup for VM 100 in Proxmox, with a deliberate schedule and retention policy, and test a restore to an isolated guest. Deployment rollback archives on the VM are not a substitute for off-host backups. Protect backup access because application state can contain credentials and notification settings.

Enable the guest's Start at boot option. Container services using restart: unless-stopped return when Docker starts unless they were intentionally stopped. Verify graceful guest shutdown through Proxmox before relying on unattended host maintenance.

After initial setup or a planned guest reboot, check:

sh
systemctl is-active docker ssh qemu-guest-agent
docker compose version
docker ps --format 'table {{.Names}}\t{{.Status}}'
df -h
getent hosts up.home.rwk.me

Verify the deployed applications through their direct and HTTPS endpoints. Check free disk space, container health, and a representative application function. Use the Proxmox console if a network or SSH change prevents remote access.