OPNsense

Home›Blog›Security & Troubleshooting›OPNsense Won't Boot: Symptom-by-Symptom Troubleshooting

OPNsense Won't Boot: Symptom-by-Symptom Troubleshooting

Category: Security & Troubleshooting · Updated 2026-11-19 · Applies to OPNsense 26.7

Illustration of a firewall that failed to start with diagnostic tools beside it

When OPNsense will not boot, the fix depends entirely on what the screen actually shows, so work by symptom: a Secure Boot refusal, a "not a bootable disk" message, a boot that stalls partway, and a system that boots but never reaches the console menu are four different problems with four different causes. Find your symptom below, apply its fix, and use its confirmation step before moving on. Changing settings at random across all four is how a recoverable system becomes a reinstall.

First, sixty seconds of observation

Before touching anything, watch one full boot attempt and note three things: the exact message on screen (photograph it if you can), the last line shown before it stops, and whether the machine beeps, restarts, or sits silently. Also note what changed recently: a firmware update, a BIOS change, new hardware, a power cut, or nothing at all. Boot failures after a change are usually the change; boot failures after a power event point at the storage. The console is your main tool throughout, so if the box is headless, attach a monitor or use the serial console at 115200 baud, as described in our serial console setup guide.

Symptom 1: Secure Boot error or the firmware refuses the loader

What you see: messages about Secure Boot policy, an unsigned or blocked bootloader, or the machine skipping the OPNsense drive entirely after a BIOS update or on new hardware.

Likely cause: Secure Boot is enabled in the machine's UEFI settings. OPNsense's own guidance for virtual environments states the same rule that applies on hardware: Secure Boot must be disabled for the system to boot. Mini PCs and second hand machines often ship with it on, and a BIOS reset can turn it back on without telling you.

Fix: enter the UEFI/BIOS setup (the key varies by vendor, commonly Del, F2, or F10 at power on), find the Secure Boot setting, and disable it. Save and reboot. While you are in there, confirm the boot mode and drive order as well, because the same screen usually holds Symptom 2's fix.

How to confirm: the machine proceeds past the firmware stage into the OPNsense loader and boot menu instead of stopping at the policy message. If disabling Secure Boot changes nothing, your symptom is one of the ones below, not this one.

Symptom 2: "Not a bootable disk" or no boot device found

What you see: the firmware reports no bootable device, boots straight into BIOS setup, or tries the network (PXE) and gives up.

Likely cause: one of three things. The boot order does not include the OPNsense drive. The drive is not detected at all (cable, slot, or a dead drive). Or the drive's boot structures are damaged, which happens after failed writes, power loss during an update, or installing to the wrong target disk. On fresh installs, a fourth cause is common: the installer was written or booted in a mode that does not match how the machine now tries to boot it.

Fix: work in order.

  1. In BIOS setup, check the storage list. If the drive does not appear, reseat it or try another slot or cable; no software fix exists for an invisible drive.
  2. If the drive appears, put it first in the boot order and disable network boot temporarily so failures are obvious rather than hidden behind PXE timeouts.
  3. If the order is right and it still refuses, boot the OPNsense installer from USB (the same live environment you installed from) and check whether the installed system is visible. Our bootable USB guide covers making that medium. From the live environment you can inspect the disk and, worst case, reinstall and import your configuration, which the installer supports through the OPNsense Importer.

How to confirm: with the USB removed, the machine boots from its own drive into OPNsense. Confirm the drive, not the USB, did the work, or you have only proved the USB boots.

Symptom 3: Boot starts, then stalls or loops

What you see: the loader and kernel messages appear, then output stops at the same line every time, the machine reboots in a loop, or the screen fills with repeated device errors.

Likely cause: a device the kernel is probing misbehaves (new NICs, USB devices, and add in cards are the usual suspects), a filesystem left dirty by power loss is waiting or failing its checks, or the system is booting but the console has switched somewhere you cannot see, for example to a serial port on a headless box. A stall at the same line across two boots is a strong hint about which device was being initialised at that moment.

Fix:

  1. Remove everything non essential: USB drives, extra cards, dongles. Boot again. Add devices back one at a time until the stall returns; the last device added is your problem.
  2. If the stall mentions the filesystem or a mount failure, let any automatic checks finish without interrupting power. Repeated hard power cuts during checks make things worse, not better.
  3. For an installed system that recently changed, boot the installer USB and use the Importer to bring your configuration into the live environment; this separates "the installation is damaged" from "the configuration triggers it" before you commit to a reinstall.
  4. If you run OPNsense in a VM and the stall began after hypervisor changes, verify the machine type and that hardware offloading expectations still match; the official VM guidance supports both i440FX and Q35 chipsets, and guest settings changed during other maintenance are a classic cause. Our Proxmox install guide lists the settings a healthy VM uses.

How to confirm: two consecutive clean boots to the console menu with all essential hardware attached. One lucky boot is not a fix.

Symptom 4: It boots, but the console menu or login never appears

What you see: boot messages scroll, then a blank screen, a frozen prompt, or output that stops just before the familiar menu with options 0 to 13.

Likely cause: the console is being delivered somewhere else. On machines installed from the serial image, the live console is the serial port at 115200 baud and the VGA screen can sit quiet; on others, a display sleeps and only looks dead. Less often, a service hangs late in boot and the menu waits behind it.

Fix: check the other console before assuming failure. Connect a serial cable and open a terminal at 115200 baud, or wake and swap the display and its cable. Press a key: a live console responds. If the menu appears on serial, the system was never broken, and you can reconfigure console preferences from the working console. If neither console answers after several minutes, treat it as Symptom 3 and remove devices, because a late hang and a hidden console look identical from one side.

How to confirm: the console menu (options 0 through 13, including 1 Assign interfaces and 8 Shell) answers on the console you actually intend to use.

Symptom 5: It broke right after an update

What you see: the system worked before a firmware update and now stalls, loops, or boots into errors, with no hardware changes in between.

Likely cause: the update itself, an interrupted update (power or network loss mid upgrade), or a plugin that did not survive the new version. Major upgrades are offline operations, and the official update guidance recommends doing them with a VGA display or serial console attached precisely so this moment is visible rather than guessed at.

Fix: do not stack more changes on top. Boot the installer USB for the current release and reinstall, using the Importer to pull in your existing configuration when offered, which preserves your settings while replacing the damaged system files. If you keep configuration backups off the box (you should; see our backup and restore guide), a clean install plus a restore achieves the same result. After it boots, review the release notes for your series before reapplying updates, and reintroduce plugins one at a time.

How to confirm: the system boots to the console menu, your interfaces and rules are present, and the firmware status page shows a coherent version with no half applied update pending.

Symptom 6: Nano or embedded system boots differently than expected

What you see: an embedded install on USB, SD, or CF media seems to lose logs and settings context after reboots, or behaves unlike a friend's full install.

Likely cause: this one is often not a fault. The nano image is a preinstalled embedded image whose /var/log and /tmp live on RAM disks by design, so log data does not survive a reboot. If the system boots and routes, it is working as designed; if it does not boot at all, treat the media as suspect, because SD and CF cards wear out, and the official sizing guidance steers anything beyond light duty toward an SSD install. Our nano image guide explains when nano is the right choice and when it is not.

Fix: for a genuinely unbootable nano system, rewrite fresh media or migrate to an SSD install and import the configuration. How to confirm: a clean boot from the new media with your configuration imported and interfaces assigned.

A note on reading boot output without fear

The scrolling text during boot looks alarming even on a healthy system: drivers probe hardware that is not there, services announce themselves in terse lines, and warnings appear that have been harmless for years. The signal is not the volume of text, it is where the text stops. A healthy boot keeps moving and ends at the console menu. A sick boot stops at the same place every time, repeats the same error, or never starts. Compare like with like: photograph a healthy boot once, when everything works, and keep it with your backups. During a failure you will be grateful for a reference that shows what normal looked like on your exact hardware, instead of a stranger's description of theirs.

Protecting the next boot before it happens

Most boot failures are survivable because someone prepared, and miserable because nobody did. Four habits cover nearly everything in this article. First, attach a console you can actually reach (a monitor that lives near the box, or a serial cable already plugged in) before you need it at midnight. Second, make updates a deliberate event: read the release notes, take the backup, and update while you are present with console access, never remotely as the last task of the day. Third, after any BIOS or hardware change, do one watched reboot immediately, while the change is fresh in your mind and easy to reverse. Fourth, keep the storage healthy: an SSD install on a reasonable power supply outlives a drawer of SD cards, and the sizing guidance exists partly because worn media fails at boot first.

If the box protects a household or a small office, also decide in advance what "down" means operationally: a spare router, a phone hotspot plan, or simply acceptance that the internet waits for the repair. That decision, made calmly now, is worth more during an outage than any single fix in the symptom lists above.

When the honest answer is reinstall

Some failures are not worth heroics: a drive that disappears from BIOS intermittently, storage errors that return after a clean install, or a system with no backup and a scrambled configuration. OPNsense is forgiving here: installation is quick, the Importer rescues configurations from an existing disk, and a downloaded backup restores the rest. The decision rule is time: if two focused passes through your symptom's fix list have not moved you forward, reinstall and import, then spend the saved hours on the backup routine that makes the next failure boring.

  • Keep a current configuration backup off the firewall itself.
  • Keep the installer USB you trust, labelled with its version.
  • After any recovery, write down the symptom and the fix while you remember them.
  • Revisit your update habits against our guide to updating OPNsense safely, because update day is when most boot problems are born.

A firewall that will not boot feels like an emergency, but it is almost always one of a handful of ordinary causes wearing a frightening costume. Match the symptom, fix that layer, and confirm the fix before you declare victory.

Keep going in Security & Troubleshooting