Next-Gen Installer, Multi-Disk RAID, Btrfs & ZFSBootMenu
Deploy unbrickable bare-metal appliances via the interactive axonwall-console --install TUI or zero-touch --unattended manifests. Choose across four unified editions—axonwall (Converged NGFW), minimal (Reduced Attack Surface), vault (Immutable S3 Backup), and k3s-storage (Barebones K3s + OpenZFS/OpenEBS)—paired with Smart NVMe/SSD/HDD tiering, LUKS2 + TPM 2.0 auto-unlock, mdadm RAID-1 A/B slots, in-kernel Btrfs + Snapper Root, or native ZFSBootMenu v3.1.0 Root-on-ZFS Boot Environments.
Next-Gen Installer TUI, Unified Editions & Smart Tiering
Booting the unified axonwall-debian13-amd64.iso launches axonwall-console --install directly on VGA and serial (ttyS0) consoles:
Express Mode inspects hardware topology and provisions optimal defaults in a single confirmation. 5-Step Custom Wizard walks operators through (1) Edition (axonwall NGFW/Converged, minimal Reduced Attack Surface, vault S3 Immutable Backup, or k3s-storage Barebones K3s + OpenZFS/OpenEBS Host), (2) Boot Mode (ab, zbm, btrfs, or rolling), (3) Smart Disk Tiering, (4) LUKS2 + TPM 2.0 Encryption, and (5) Initial Management NIC & Credentials.
Automatically classifies block devices by rotational media, NVMe namespace speed, and capacity—assigning fast NVMe devices to OS boot mirrors, dedicated ZFS SLOG (logDisks), and L2ARC (cacheDisks), while grouping bulk SAS/SATA drives into RAID-Z/dRAID dataDisks and spareDisks.
Optional full-disk/pool at-rest encryption formats backing partitions with AES-256-XTS LUKS2 and seals a hardware recovery key into the onboard TPM 2.0 via systemd-cryptenroll bound to PCRs 0 + 2 + 7 (UEFI firmware, Option ROMs, and Secure Boot state), enabling unattended reboots that fail closed if hardware or boot chain is tampered with.
Automate mass fleet rollouts via axonwall-console --install --unattended <config.yaml|https://...>, or generate a bootable NoCloud seed image with axonwallctl cloudinit export --iso -o cidata.iso.
# Launch interactive TUI installer (Express 1-Click or 5-Step Custom Wizard) sudo axonwall-console --install # Automated zero-touch installation from local YAML or HTTPS provisioning endpoint sudo axonwall-console --install --unattended https://provision.corp.internal/fw-branch-04.yaml # Export a NoCloud seed ISO (cidata.iso) from an existing or staged configuration axonwallctl cloudinit export --iso -o cidata.iso
System Disk Mirroring vs. Dedicated ZFS Storage
When installing Axonwall on servers with two or more disks, administrators can mirror the core operating system across disks using hardware or kernel RAID-1, while partitioning the remaining disk capacity for OpenZFS:
Mirrors the entire Axonwall OS layout across 2+ member disks. Each disk contains cloned UEFI ESP partitions, Slot A (axon_root_a on /dev/md0), Slot B (axon_root_b on /dev/md1), Swap (/dev/md2), and Persistent Data (/dev/md3).
- GRUB bootloader installed on every mirror member
- Survives complete drive loss without loss of firmware watchdog
- Automatic array sync with write-intent bitmaps
Instead of consuming the whole disk for the root filesystem, the installer lets you carve out a dedicated ZFS storage slice labeled axon_zfs_1 (or across all member disks in RAID setups).
- Specify fixed gigabytes (e.g. 500 GB) or -1 for all remaining disk space
- Select ZFS RAID topology: Stripe, Mirror, RAID-Z1, RAID-Z2, RAID-Z3, or dRAID
- Bootstrap spec saved to
/var/lib/axonwall/install-zfs.json
Step-by-Step Installer Walkthrough
1. System Disk Selection & Mirror Mode
The installer scans the SCSI/NVMe bus and displays all discovered physical drives. If multiple drives are present, you are prompted:
Found physical storage disks: [1] /dev/sda (Samsung SSD 990 PRO 2TB - 2000 GB) [2] /dev/sdb (Samsung SSD 990 PRO 2TB - 2000 GB) Select system installation target: (1) Single Disk: Install Axonwall on /dev/sda (2) Multi-Disk Mirror: Configure mdadm RAID-1 across /dev/sda and /dev/sdb Enter choice [1-2] (default: 1): 2
2. A/B Dual-Root Partition Sizing
Choose the firmware slot size for the A/B root filesystems (Standard: 16 GB, Compact: 8 GB, Expanded: 32 GB, or Custom).
3. Dedicated ZFS Storage Partition Allocation
Specify how much space on the system drive(s) to allocate for OpenZFS NAS pools:
Configure dedicated ZFS storage partition on system disk(s)? - Enter '0' to skip dedicated ZFS partition on system disk - Enter size in GB (e.g. '500') to reserve a fixed slice - Enter '-1' to allocate all remaining disk space for ZFS storage ZFS Partition Size (GB, 0 to skip, -1 for all remaining) [default: 0]: -1 Select ZFS Storage Pool Topology: (1) stripe - Single VDEV / maximum capacity, no parity (2) mirror - 2-way mirror (RAID-1) across disk partitions (3) raidz1 - RAID-Z1 (Single parity, requires 3+ disks) (4) raidz2 - RAID-Z2 (Dual parity, requires 4+ disks) (5) raidz3 - RAID-Z3 (Triple parity, requires 5+ disks) (6) draid1 - Distributed Spare RAID-1 (7) draid2 - Distributed Spare RAID-2 Enter ZFS topology [1-7] (default: 2): 2 Enter ZFS pool name [default: tank]: tank
/var/lib/axonwall/install-zfs.json. On first boot, the privileged Go daemon (axonwalld) automatically executes zpool create with optimal ashift=12, LZ4 compression, POSIX ACLs, and mounts /mnt/tank with zero manual CLI intervention.GPT Partition Structure (Single vs. Mirrored A/B)
Drive /dev/sda (and /dev/sdb if RAID-1): +-------------------------------------------------------------------------------+ | Part | Label | Size | Filesystem | Purpose | +-------------------------------------------------------------------------------+ | p1 | EFI System | 512 MB | FAT32 (vfat) | Cloned UEFI ESP + GRUB | | p2 | axon_root_a | 16 GB | ext4 (or md0) | Firmware Slot A (Active) | | p3 | axon_root_b | 16 GB | ext4 (or md1) | Firmware Slot B (Standby) | | p4 | axon_swap | 4 GB | swap (or md2) | Linux Kernel Swap | | p5 | axon_data | 32 GB | ext4 (or md3) | Persistent /var/lib/axonwall| | p6 | axon_zfs_1 | Rest | ZFS Member | OpenZFS Storage Pool (tank) | +-------------------------------------------------------------------------------+
ZFSBootMenu Root-on-ZFS Mode (v3.1.0)
As an alternative to traditional ext4 A/B dual-root slots, Axonwall provides native ZFSBootMenu v3.1.0 (ZBM) support (installMode: "zbm"). Under ZBM, the entire operating system boots directly from an OpenZFS root pool (zroot) with native support for any zpool topology (stripe, mirror, raidz1/2/3, draid1/2/3).
System root datasets reside in zroot/ROOT/<be_name>. OTA firmware updates snapshot the active BE (e.g. zroot/ROOT/axonwall-0.12.4@pre-update-v0.13.0) and clone it instantaneously with zero wasted disk space into a new target BE (zroot/ROOT/axonwall-0.13.0).
- Instantaneous atomic updates and rollbacks via
axonwallctl zbm - Multiple historical Boot Environments retained simultaneously
- Non-root data isolated in
zroot/data
ZFSBootMenu hooks execute before the Linux kernel is loaded. Hook 20-axonwall-watchdog.sh inspects org.axonwall:bootcount and org.axonwall:bootlimit=3.
- Reverts
bootfstoorg.axonwall:fallback_beon 3 failures - Autonomous promotion (
zfs promote) upon verified daemon quorum - Full Web Console UI under System > Updates and TUI tab
[u] ZBM/Boot
Physical Disks (Single, Mirror, RAID-Z, or dRAID pool):
+-------------------------------------------------------------------------------+
| Part | Label | Size | Filesystem | Purpose |
+-------------------------------------------------------------------------------+
| p1 | EFI System | 1024 MB | FAT32 (vfat) | UEFI ESP: EFI/ZBM/VMLINUZ.EFI|
| p2 | axon_swap | 4 GB | swap | Linux Kernel Swap (Optional)|
| p3 | axon_zbm_root | Rest | OpenZFS VDEV | zroot (Root Pool VDEV member)|
+-------------------------------------------------------------------------------+
Dataset Hierarchy:
zroot
├── ROOT
│ ├── axonwall-0.12.4 (mountpoint=/, org.zfsbootmenu:commandline="...")
│ │ └── @pre-update-v0.13.0 (Point-in-time snapshot)
│ └── axonwall-0.13.0 (mountpoint=/, clone/promoted next-boot BE)
└── data
└── axonwall (mountpoint=/var/lib/axonwall, persistent config & logs)# List all ZFSBootMenu Boot Environments, active bootfs, and snapshots axonwallctl zbm list # Create a point-in-time snapshot, clone into a new BE, and activate for next boot sudo axonwallctl zbm snapshot axonwall-0.13.0 pre-change sudo axonwallctl zbm clone axonwall-0.13.0@pre-change axonwall-canary sudo axonwallctl zbm activate axonwall-canary # Prune obsolete Boot Environments or snapshots sudo axonwallctl zbm delete axonwall-0.12.2 sudo axonwallctl zbm delete-snapshot axonwall-0.13.0@pre-change # Stage cryptographically verified OTA firmware update (SHA-256 + OpenSSL .sig # verified against /etc/axonwall/release-signing.pub before writing standby slot/BE) sudo axonwallctl update --version 0.13.0 sudo axonwallctl update --version 0.13.0 --slot B sudo axonwallctl update --btrfs --version 0.13.0 # Btrfs + Snapper pre/post mode sudo axonwallctl update --apt # Rolling single-root APT upgrade mode
ZFSBootMenu Visual Theming & Mirrored ESP Sync
Axonwall provides full visual theming for ZFSBootMenu v3.1.0 across the UEFI GOP graphics splash screen, console typography, and 5 tactical ANSI color palettes—automatically synchronizing the updated UKI across all mirrored EFI System Partitions (/boot/efi, /boot/efi2, etc.).
A 24-bit uncompressed Windows BMP (640x480) is embedded directly into the Unified Kernel Image (UKI) .splash section via systemd-stub, rendering seamlessly before kernel initialization.
On 4K and Retina consoles, small text is enlarged using rd.vconsole.font=ter-132b (32px bold Terminus) with automatic fallback (zbm.autosize=1).
Early pre-boot hooks (15-axonwall-theme.sh) program hardware VGA registers with five built-in palettes: cyber-cyan, emerald-ops, amber-tactical, crimson-soc, and mono-stealth.
# Check current theme, embedded UKI splash section, mirrored ESPs, and active font axonwall-zbm-theme status # Set font for 4K / High-DPI displays (ter-132b, ter-v32b, ter-v28b, ter-v24b, or default) sudo axonwall-zbm-theme set-font ter-132b # Switch 16-color ANSI console palette (cyber-cyan, emerald-ops, amber-tactical, crimson-soc, mono-stealth) sudo axonwall-zbm-theme set-palette emerald-ops # Install custom splash graphic (automatically converts PNG/JPG to uncompressed BMP) sudo axonwall-zbm-theme set-splash /root/custom-brand.png # Re-embed UKI .splash section and sync EFI/ZBM/VMLINUZ.EFI across all mirrored ESP partitions sudo axonwall-zbm-theme apply
Btrfs + Snapper Root Mode (installMode: "btrfs")
For systems seeking a lightweight, in-kernel Copy-on-Write alternative to ZFSBootMenu—without external DKMS modules—select Btrfs + Snapper Root (installMode: "btrfs"). Single-disk targets use standard Btrfs with DUP metadata and compress=zstd:1, while multi-disk system targets automatically provision native Btrfs RAID-1 (-d raid1 -m raid1) across system drives.
The active OS root resides on subvolume @, while snapshots, logs, and persistent appliance state live on sibling subvolumes so rolling back the OS never erases audit logs, snapshots, or configuration:
@→/(Root filesystem)@snapshots→/.snapshots(Snapper timelines & pre/post pairs)@var_log→/var/log(Audit & syslog persistence)@var_lib_axonwall→/var/lib/axonwall(or LUKS2axon_data)
Every firmware update (axonwall-update --btrfs) and Kernel Studio installation automatically takes a pre-change Snapper snapshot, applies the change onto @, takes a paired post snapshot, and regenerates /boot/grub/btrfs-snapshots.cfg.
- Boot any historical read-only snapshot from GRUB submenu Axonwall Btrfs Snapshots
- Atomic subvolume promotion via Web UI,
axonwall-tui, oraxonwall-btrfs-grub rollback <id> - Optional dedicated extra disks still provision OpenZFS pool
tank
# GPT Partition Layout (Single or Dual-Disk Native Btrfs RAID-1): # p1: 1MiB-4MiB bios_grub (Legacy BIOS GRUB support) # p2: 4MiB-516MiB AXON_EFI (512 MiB FAT32 ESP, mirrored across system disks) # p3: [optional] axon_data (4096 MiB LUKS2 AES-XTS-512 when encryption enabled) # pN: remaining axon_btrfs (Single Btrfs volume or native RAID-1 mirror) # Create manual checkpoint, refresh GRUB menu, or atomically roll back to Snapshot #1 sudo snapper -c root create --description "Pre-maintenance checkpoint" sudo axonwall-btrfs-grub sync sudo axonwall-btrfs-grub rollback 1
Minimal Reduced-Attack-Surface Edition (minimal)
When minimal attack surface matters most, select the Minimal Edition (minimal) in the GRUB2/ISOLINUX boot menu, axonwall-console --install wizard, Debian d-i preseed (preseed-minimal.cfg), Cloud-Init schema, or install axonwall-minimal_0.13.0_amd64.deb directly. It provisions only the core boot, management, and firewall baseline while letting operators add capabilities on demand later.
- Boots cleanly with minimal package footprint; optional packages and workstation extras are purged on target
- DHCP, DNS, OpenZFS NAS, and KVM/LXC roles remain disabled unless explicitly enabled
- Cleanly falls back to
/bin/bashacross OpenLDAP/SSSD directory templates whenzshis absent - Dedicated updater track via
/etc/axonwall/editionacross A/B, ZFSBootMenu, Btrfs + Snapper, and Rolling slots
Whenever you later need Kubernetes (core-k3s), OpenZFS NAS (core-nas), workstation shells (cli-workstation), or GNU Emacs/Rosenpass, enable or install them on demand from System → Plugins (/system/plugins), /v1/system/plugins/action, or axonwallctl plugin install <id>. Plugins with SHA-256-verified bundled binaries self-stage into isolated bin/ trees automatically.
Barebones K3s + OpenZFS/OpenEBS Storage Host (k3s-storage)
Dedicated compute and storage nodes do not need NGFW, Suricata IDS, or Samba/NFS appliances enabled. Both the Debian 13 and Alpine Linux ISOs expose the k3s-storage profile directly in UEFI GRUB / BIOS ISOLINUX boot menus as well as Step 1 of axonwall-console --install.
core-k3s&core-nasenabled cleanly in host mode (core-firewall=false)- Pre-staged OpenZFS ZFS-LocalPV CSI operator (
/var/lib/rancher/k3s/server/manifests/zfs-operator.yaml) backed by your selected RAID pool - First-boot reconcilers (
axonwall-k3s-firstboot,axonwall-k3s-verify,axonwall-k3s-status) plus hardened minimal hostnftables.conf& sysctl tuning - Alpine Linux OpenRC parity via
axonwall-firstboot-alpine+k3s/k3s-agentinit scripts
During installation (or via unattended k3sStorage.role in YAML / Cloud-Init), designate the host as either an initial control-plane server node or a agent storage worker joining an existing cluster URL with a pre-shared join token. Pair with Custom Hardened Kernels to lock down ZFS + K3s host kernels further.