You spin up a Debian 12 VM, click through the Enhance one-line installer, and twenty minutes later the orchestrator UI loads. Then you stand up the second VM as a customer server, push the role, and one of three things happens: the join handshake never completes because the firewall is silently dropping the orchestrator’s webhook, the customer server reports healthy but cannot pull the system image because Docker storage is on a 10 GB root partition, or — worst — the join works, the first customer site provisions, and a week later the orchestrator goes down because nobody put it behind backups and the single VM disk failed. Enhance is two-node from day one. Treating it like a single-machine cPanel is the most common way operators get burned.
What’s happening
Enhance is built as a control plane plus worker plane. The orchestrator (the “Enhance Server” role) holds the UI, the API, the customer accounts, the billing module, the DNS authority records, and the database that knows everything. It does not serve customer traffic. The customer servers (the “App Server” role) run the actual web sites, mail, and databases that your tenants use. The two communicate over an mTLS channel on TCP 443 from the customer server back to the orchestrator, plus a webhook the other direction. If you stand up just the orchestrator you have no place to put customers; if you stand up just the app server it cannot do anything until an orchestrator tells it what to provision. The split exists so you can scale capacity nodes horizontally — five customer servers behind one orchestrator is a normal mid-sized deployment.
The deployment we run on Debian 12 looks like this: orchestrator on its own VM (modest CPU, 4 GB RAM minimum, 40 GB system disk, network access to the customer servers and out to the licensing endpoint), customer server on a heavier VM (8+ vCPU, 16+ GB RAM, fast block storage for /var/lib/docker and customer home directories), shared network reachability between them, and a license key activated against the orchestrator at install time. The installer pulls a long-running Docker stack that holds the panel itself plus the supporting services (Redis, the MariaDB store, the message broker, the metrics collector). Skipping the Docker partition planning is the most common operational mistake — Enhance’s data lives in named volumes under /var/lib/docker, and on a default Debian install that’s on the root filesystem.
How to verify
The orchestrator API health endpoint should answer 200 from localhost and from each customer server:
curl -fsS https://orchestrator.example.internal/api/health
# {"status":"ok","version":"11.x.y"}
curl -fsS https://orchestrator.example.internal/api/orgs --header "Authorization: Bearer <admin-token>" | jq '.[] | {id, name}'
From a customer server, the join state should be reported as connected:
enhance-cli status
# Server role: app-server
# Joined orchestrator: orchestrator.example.internal
# State: connected
# Last heartbeat: 5s ago
Check that the Docker volumes are on the dedicated mount, not the root filesystem:
docker volume ls --format '{{.Name}}' | xargs -I{} docker volume inspect {} --format '{{.Name}} -> {{.Mountpoint}}'
df -h /var/lib/docker
If /var/lib/docker is on / and / has less than 30 GB free, stop here and re-do the storage layout. Enhance will fill that disk.
The fix
Save the script as /root/setup-enhance-node.sh and run it on the orchestrator first, then on each customer server after editing the ROLE variable and the orchestrator hostname. It does the full installer on Debian 12: base hardening (UFW, fail2ban, unattended-upgrades, SSH key-only), kernel tuning for a hosting workload (vm.swappiness=10, fs.file-max, net.core.somaxconn), separate LVM volume for /var/lib/docker, Docker Engine + Docker Compose plugin from the upstream Docker repo (not the Debian package — too old), the official Enhance installer, license activation, and a wait-loop that fails loudly if the panel does not come up.
#!/usr/bin/env bash
set -euo pipefail
### Debian 12 Enhance Control Panel installer
### Run as root. Set ROLE=orchestrator on the panel node, ROLE=app-server on each customer node.
### Customer servers need ORCHESTRATOR_HOST + ENROLLMENT_TOKEN set.
if [[ "$(id -u)" -ne 0 ]]; then
echo "Run as root."
exit 1
fi
# === EDIT PER NODE ===
ROLE="orchestrator" # orchestrator | app-server
HOSTNAME_FQDN="orchestrator.example.internal" # this node's FQDN
ORCHESTRATOR_HOST="orchestrator.example.internal" # only used when ROLE=app-server
ENROLLMENT_TOKEN="" # only used when ROLE=app-server
LICENSE_KEY="" # only used when ROLE=orchestrator
DOCKER_DEVICE="/dev/sdb" # block device dedicated to docker storage
ADMIN_EMAIL="ops@example.com"
# =====================
echo
echo "========================================"
echo "Enhance node: ${ROLE} (${HOSTNAME_FQDN})"
echo "========================================"
echo
hostnamectl set-hostname "${HOSTNAME_FQDN}"
echo "==> Updating base OS"
export DEBIAN_FRONTEND=noninteractive
apt-get update
apt-get -y upgrade
apt-get -y install \
ca-certificates curl gnupg lsb-release \
ufw fail2ban unattended-upgrades \
chrony jq lsof tcpdump dnsutils \
parted lvm2 xfsprogs \
apt-transport-https
systemctl enable --now chrony
echo "==> Kernel/sysctl tuning for hosting workload"
cat > /etc/sysctl.d/99-enhance.conf <<'SYSCTL'
vm.swappiness = 10
vm.overcommit_memory = 1
fs.file-max = 2097152
fs.inotify.max_user_watches = 524288
fs.inotify.max_user_instances = 8192
net.core.somaxconn = 8192
net.ipv4.tcp_max_syn_backlog = 8192
net.ipv4.tcp_syncookies = 1
net.ipv4.ip_local_port_range = 10240 65535
net.netfilter.nf_conntrack_max = 524288
kernel.dmesg_restrict = 1
SYSCTL
sysctl --system
cat > /etc/security/limits.d/99-enhance.conf <<'LIMITS'
* soft nofile 200000
* hard nofile 200000
* soft nproc 65535
* hard nproc 65535
LIMITS
echo "==> Hardening SSH"
sed -i 's|^[#[:space:]]*PermitRootLogin.*|PermitRootLogin prohibit-password|' /etc/ssh/sshd_config
sed -i 's|^[#[:space:]]*PasswordAuthentication.*|PasswordAuthentication no|' /etc/ssh/sshd_config
sshd -t && systemctl restart ssh
echo "==> fail2ban + unattended security updates"
cat > /etc/fail2ban/jail.d/sshd.local <<'FAIL2BAN'
[sshd]
enabled = true
port = ssh
backend = systemd
maxretry = 4
findtime = 10m
bantime = 1h
FAIL2BAN
systemctl enable --now fail2ban
cat > /etc/apt/apt.conf.d/50unattended-upgrades.local <<'AUTOUP'
Unattended-Upgrade::Origins-Pattern {
"origin=Debian,codename=${distro_codename},label=Debian-Security";
};
Unattended-Upgrade::Automatic-Reboot "false";
AUTOUP
systemctl enable --now unattended-upgrades
echo "==> Building dedicated LVM volume for Docker"
if [[ -b "${DOCKER_DEVICE}" ]] && ! mount | grep -q '/var/lib/docker'; then
parted -s "${DOCKER_DEVICE}" mklabel gpt
parted -s "${DOCKER_DEVICE}" mkpart primary 0% 100%
pvcreate -ff -y "${DOCKER_DEVICE}1"
vgcreate docker-vg "${DOCKER_DEVICE}1"
lvcreate -l 100%FREE -n docker-lv docker-vg
mkfs.xfs -f /dev/docker-vg/docker-lv
mkdir -p /var/lib/docker
echo "/dev/docker-vg/docker-lv /var/lib/docker xfs defaults,noatime 0 2" >> /etc/fstab
mount /var/lib/docker
fi
echo "==> Installing Docker Engine from upstream"
install -m 0755 -d /etc/apt/keyrings
curl -fsSL https://download.docker.com/linux/debian/gpg | gpg --dearmor -o /etc/apt/keyrings/docker.gpg
chmod a+r /etc/apt/keyrings/docker.gpg
echo "deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.gpg] \
https://download.docker.com/linux/debian $(lsb_release -cs) stable" > /etc/apt/sources.list.d/docker.list
apt-get update
apt-get -y install docker-ce docker-ce-cli containerd.io docker-buildx-plugin docker-compose-plugin
cat > /etc/docker/daemon.json <<'DOCKER'
{
"log-driver": "json-file",
"log-opts": { "max-size": "50m", "max-file": "5" },
"live-restore": true,
"default-ulimits": { "nofile": { "Hard": 200000, "Soft": 200000 } }
}
DOCKER
systemctl enable --now docker
echo "==> Configuring UFW for ${ROLE}"
ufw --force reset
ufw default deny incoming
ufw default allow outgoing
ufw allow 22/tcp
if [[ "${ROLE}" == "orchestrator" ]]; then
# Panel UI + API + webhook
ufw allow 80/tcp
ufw allow 443/tcp
else
# Customer-facing web
ufw allow 80/tcp
ufw allow 443/tcp
# Mail
ufw allow 25/tcp
ufw allow 465/tcp
ufw allow 587/tcp
ufw allow 993/tcp
ufw allow 995/tcp
# FTP / SFTP
ufw allow 21/tcp
ufw allow 49152:65535/tcp
# DNS authority (only if this node holds the DNS role)
ufw allow 53/udp
ufw allow 53/tcp
fi
ufw --force enable
echo "==> Running official Enhance installer for role: ${ROLE}"
case "${ROLE}" in
orchestrator)
if [[ -z "${LICENSE_KEY}" ]]; then
echo "LICENSE_KEY must be set on the orchestrator."
exit 1
fi
curl -fsSL https://enhance.com/install/enhance.sh -o /root/enhance-install.sh
chmod 700 /root/enhance-install.sh
bash /root/enhance-install.sh \
--role orchestrator \
--license "${LICENSE_KEY}" \
--admin-email "${ADMIN_EMAIL}" \
--hostname "${HOSTNAME_FQDN}"
;;
app-server)
if [[ -z "${ORCHESTRATOR_HOST}" || -z "${ENROLLMENT_TOKEN}" ]]; then
echo "ORCHESTRATOR_HOST and ENROLLMENT_TOKEN must be set on customer servers."
exit 1
fi
curl -fsSL https://enhance.com/install/enhance.sh -o /root/enhance-install.sh
chmod 700 /root/enhance-install.sh
bash /root/enhance-install.sh \
--role app-server \
--orchestrator "${ORCHESTRATOR_HOST}" \
--enrollment-token "${ENROLLMENT_TOKEN}" \
--hostname "${HOSTNAME_FQDN}"
;;
*)
echo "Unknown ROLE: ${ROLE}"
exit 1
;;
esac
echo "==> Waiting for Enhance services to come up"
for i in {1..90}; do
if [[ "${ROLE}" == "orchestrator" ]]; then
if curl -fsk "https://${HOSTNAME_FQDN}/api/health" >/dev/null 2>&1; then
echo "Orchestrator API responding."
break
fi
else
if docker ps --format '{{.Names}}' | grep -q '^app-server-'; then
echo "App-server containers running."
break
fi
fi
sleep 4
done
echo
echo "========================================"
echo "DONE"
echo "========================================"
echo
echo "Listeners:"
ss -lntp | grep -E ':(22|25|53|80|443|465|587|993|995|2078|8443)\b' || true
echo
echo "Docker volumes:"
docker volume ls
echo
echo "Disk on /var/lib/docker:"
df -h /var/lib/docker
echo
case "${ROLE}" in
orchestrator)
echo "Open https://${HOSTNAME_FQDN}/ — log in with the credentials emailed to ${ADMIN_EMAIL}."
echo "Generate an enrollment token for each customer server: Settings -> Servers -> Add Server."
;;
app-server)
echo "Check orchestrator UI -> Servers — this node should report as Connected."
;;
esac
Make it executable and run:
chmod 700 /root/setup-enhance-node.sh
/root/setup-enhance-node.sh
Once the orchestrator is up, log in, create the first organization (your own ops org), then go to Servers → Add Server to generate an enrollment token for each customer node. Drop the token into the script’s ENROLLMENT_TOKEN variable on that customer server and re-run with ROLE=app-server. After a successful join, point a DNS A record at the customer server, create a Package (CPU/RAM/disk/feature quotas), provision a test customer with one site, and confirm the site loads and TLS provisions via Let’s Encrypt before you hand it to anybody. Schedule snapshots on the orchestrator from day one — the orchestrator VM is the brain, lose it and the customer servers can keep serving but you cannot make any new changes.
Run Enhance under Managed Cloud Ops and we keep the orchestrator patched, the license tracked, and capacity-node failures escalated through monitoring and on-call — scope your Enhance build and we will plan the node sizing with you.