0x0000007B INACCESSIBLE_BOOT_DEVICE
This one happens during startup, before Windows is running properly - which means no dump, often no event log entry, and no desktop to troubleshoot from. Windows lost access to the partition it boots from. The good news is that the most frequent cause is a firmware setting rather than a dead drive, and reverting it takes two minutes.
What you are looking at
Your PC ran into a problem and needs to restart.
Stop code: INACCESSIBLE_BOOT_DEVICE
What the error actually means
Booting Windows is a chain. Firmware hands control to the boot loader, which loads the kernel and a minimal set of drivers - crucially including the driver for whatever controller the system drive is attached to. That driver brings up the storage device, the file system driver mounts the partition, and startup continues.
Bugcheck 0x7B means that chain broke. Either the storage driver failed to initialise the device, or the file system did not recognise what it found on it. Microsoft's own description is that the operating system lost access to the system partition during startup.
The reason a firmware setting is such a common cause is that it changes which driver is needed. Windows only preloads drivers for the configuration it was installed under - switch the controller from AHCI to RAID after the fact and the required driver is not there, so the drive is simply invisible.
Reading the parameters
Only parameter 1 carries information, and interpreting it requires a debugger - which is rarely available for a machine that will not boot. Practical diagnosis here comes from the circumstances, not the parameters.
| Parameter | What it holds |
|---|---|
| 1 | Either the address of a UNICODE_STRING structure, or the address of the device object that could not be mounted. Microsoft's method for telling them apart: look at the first USHORT at that address. If it is 0x3, parameter 1 points at a device object - meaning the file system failed to initialise or did not recognise the data as a file system structure. If it is even, it is a Unicode string containing the ARC name of the device the boot was attempted from. |
| 2 | Zero. |
| 3 | Zero. |
| 4 | Zero. |
What actually triggers it
Ordered by what actually causes this in practice. The first entry is responsible for a large share of cases and is free to reverse.
| Cause | How often | Detail |
|---|---|---|
| A UEFI storage mode change | Common | Switching the SATA controller between AHCI, RAID and legacy IDE modes after Windows is installed. Microsoft explicitly lists this - changing the controller mode from legacy to AHCI in UEFI - as a cause and advises reverting it. This also happens accidentally: a firmware update or a cleared CMOS can reset the mode without you touching it. |
| A new or changed disk controller | Common | Installing a storage controller card, moving the drive to a different port type, or cloning to a machine with a different chipset. Microsoft notes some disk controllers require additional drivers to be present when Windows starts. |
| Failing storage | Common | Microsoft states plainly that this error can occur when the storage hardware has failed and cannot respond to Windows. A drive that has developed bad sectors in the boot region, or an SSD whose controller has stopped answering, produces exactly this. |
| Corrupt boot configuration or file system | Occasional | A damaged BCD store, a corrupted partition table, or file system damage from an unclean shutdown. Repartitioning the system disk is listed by Microsoft as a cause. |
| A bad update or driver change | Occasional | A storage driver update or a Windows update that fails partway can leave the boot-critical driver set inconsistent. |
| Loose or failing cabling | Occasional | Mundane and genuinely common on desktops. A SATA cable that has worked loose, or a power connector not fully seated, presents as a drive that is not there. |
| Malware affecting the boot sector | Rare | Microsoft's guidance recommends a virus scan that specifically checks the Master Boot Record. |
Narrowing it down
Because the machine will not boot, work from what changed rather than from diagnostics.
Ask what changed immediately before
Did you enter UEFI for any reason? Update firmware? Clear CMOS or replace the board battery? Add or move a drive? Install a controller card? Clone the system to new hardware? Install a Windows update?
Any firmware or storage change immediately preceding the failure is almost certainly the cause, and reverting it is the fix. This single question resolves more of these than every other step combined.
Check the storage mode in UEFI
Enter firmware setup - usually Delete, F2 or Escape at power on - and find the SATA or storage configuration. Note whether it is set to AHCI, RAID or IDE.
If it is not what Windows was installed under, change it back. If you do not know what it was installed under, AHCI is the overwhelmingly likely answer on any modern system - try that first.
Confirm the drive is visible at all
While in UEFI, check whether the boot device appears in the drive list and in the boot order.
A drive that firmware cannot see is a hardware or cabling problem, not a Windows problem - and no amount of recovery-environment work will help until it is visible. A drive that firmware sees but Windows cannot mount is a driver or file system problem.
Test whether Safe Mode works
Interrupt the boot three times to reach recovery, then Troubleshoot, Advanced options, Startup Settings, Restart, and choose Safe Mode.
Microsoft notes Safe Mode loads a core set of storage drivers that can allow the storage system to be accessed again. If Safe Mode works, the hardware is fine and a driver or configuration problem is at fault - which is a much better position to be in.
Fixes, cheapest first
Try these in order. The first is free and reverses the most common cause.
Revert the UEFI storage mode
Do you have BitLocker or Device Encryption?Changing storage-controller settings in firmware triggers a BitLocker recovery-key prompt on the next boot. If you do not have the 48-digit key, the volume is unrecoverable - which is a far worse outcome than the stop code you are trying to fix.
Windows 11 enables Device Encryption by default on much consumer hardware, and many people do not know it is on. If the machine still boots into Windows at all, check first:
manage-bde -statusIf it reports protection on, sign in to account.microsoft.com/devices/recoverykey from any other device and save the key before you touch firmware. If the machine no longer boots, retrieve the key before changing anything - you may need it either way.
In firmware setup, set the SATA or storage controller mode back to what it was - in practice, almost always AHCI. Save and exit.
If you genuinely need to change modes on a working installation, that has to be prepared for from inside Windows before rebooting. Changing it blind is what produces this stop code.
Watch outThis is the same territory as the storage-driver swap on DPC_WATCHDOG_VIOLATION. Changing storage controller configuration on a system using RAID or Intel VMD can make it unbootable - and VMD is enabled by default on many recent Intel laptops with no RAID array present, so check the actual mode rather than assuming.
Reseat cables and drives
Power down fully and unplug at the wall. Reseat both ends of the SATA data cable and the power connector, or reseat the M.2 drive and its retaining screw. Try a different SATA port and a different cable if you have one.
Unremarkable, and it fixes a real share of desktop cases.
Run Startup Repair
From the recovery environment, choose Troubleshoot, then Advanced options, then Startup Repair. It rebuilds boot configuration automatically and handles the simpler corruption cases without you needing to know what is wrong.
- Interrupt boot three times with the power button to reach recovery.
- Troubleshoot, Advanced options, Startup Repair.
- Let it complete, then attempt a normal boot.
Check the disk from recovery
Open Troubleshoot, Advanced options, Command Prompt. Drive letters in recovery are usually not what you expect, so establish them first.
- Run
diskpart, thenlist volume, thenexit. Note which letter holds Windows. - Run
chkdsk X: /f /ragainst that volume. - Then
sfc /scannow /offbootdir=X:\ /offwindir=X:\Windowsusing the same letter.
Watch outIf the drive is suspected of failing,
/rcan accelerate the failure. Image the drive first if the data matters.- Run
Roll back the update or restore
If this began after an update, use Troubleshoot, Advanced options, Uninstall Updates - it offers the latest quality update and the latest feature update separately.
Failing that, System Restore under the same menu takes the machine back to a restore point from before the problem, which reverses driver and configuration changes without touching your files.
Update firmware, or check controller compatibility
Microsoft's guidance suggests checking for updates to system UEFI and storage controller firmware, and confirming that the disk controller is compatible with the installed Windows version. Relevant particularly after adding a controller card or moving a drive into different hardware.
SSD firmware updates that fix drives disappearing from the host are real and have been shipped by several vendors.
Watch outFirmware updates on a machine that will not boot usually have to be done from a USB stick using the vendor's own tool. Do not interrupt one.
When it is the hardware, not Windows
Signals that the drive has genuinely failed rather than being misconfigured:
- The drive does not appear in UEFI at all, or appears intermittently between boots.
- It appears with the wrong capacity or a garbled model name.
- Booting worked yesterday with no firmware or hardware change in between.
- The machine had been freezing during file operations before this started.
- A different drive boots fine in the same port with the same cable.
- SMART data - readable by connecting the drive to another machine - shows reallocated sectors or degraded health.
If the drive is failing, recovering data comes before recovering Windows. Connect it to another machine as a secondary drive, or boot that machine from a Linux USB stick, and copy what matters off before running repair tools. Repair operations write to a surface that is already unreliable, and every additional write is a chance to lose something you wanted.
Frequently asked
I changed the SATA mode in BIOS and now it will not boot. Can I fix it without reinstalling?
Yes - change it back. That is the entire fix in most cases, and Microsoft's own guidance says to revert such changes. If you specifically need the new mode, the switch has to be prepared from inside a working Windows first so the required driver is loaded at boot; done blind, it produces exactly this stop code. Set it back, boot normally, and decide from there.
Windows will not boot at all. Where do I even start?
Force the recovery environment: power on, then hold the power button to kill the machine as soon as you see the Windows logo. Do that three times and the fourth boot goes into recovery. From there Startup Repair is the low-effort first attempt and Command Prompt gives you diskpart and chkdsk. If recovery itself will not load, boot from Windows installation media and choose Repair your computer.
Why is there no dump file or event log entry?
Because the failure happens before Windows has started the subsystems that write them. Microsoft notes the error logs might be empty because the OS has not booted far enough to start those subsystems. That is why diagnosis here relies on what changed and on firmware-level checks rather than on the usual dump analysis.
I cloned my drive to a new SSD and now I get this. What went wrong?
Usually one of three things: the clone did not include the EFI system partition or boot configuration, the new drive is on a different controller type than the original, or the boot order still points at the old drive. Check the boot order first since it is free. Then confirm the storage mode matches. Startup Repair from installation media handles most remaining clone cases by rebuilding the boot configuration for the new drive.
Related errors
- 0x000000EFCRITICAL_PROCESS_DIEDA process Windows can't live without exited. Often a boot loop, so the fixes run from recovery.
- 0x00000133DPC_WATCHDOG_VIOLATIONA driver hogged the CPU past a hard deadline. Storage firmware and SATA drivers cause most of them.
- 0x00000154UNEXPECTED_STORE_EXCEPTIONMemory compression hit an error it couldn't handle. Check the drive before the RAM.