Rescue mode

Rescue mode boots your instance from the VirtConsole Rescue image, a Debian 13 live system with your instance’s own disks attached but not mounted. You can use it to repair file systems, reset forgotten passwords, fix boot problems, or access files that the operating system cannot boot into.

Enable rescue mode

Rescue mode is available only for KVM instances that are deployed but not suspended or currently running a task. To enter rescue mode:

1. In the user panel go to Compute > Instances and click Manage next to your instance.
2. Click Rescue mode in the header section.
3. Click Enter rescue mode. The instance will power cycle and boot into the rescue image.

Your rescue mode is active when you see the red Rescue mode banner at the top of the instance page, showing your root username, IP addresses, and a temporary password.

API

To enable rescue mode programmatically:

POST /api/instance/{id}/rescue
Content-Type: application/json

{
  "enable": true
}

The response includes the temporary root password and SSH details:

{
  "success": true,
  "rescue": {
    "active": true,
    "since": "2026-09-24T10:30:00Z",
    "username": "root",
    "password": "TemporaryPassword123"
  }
}

Access your instance in rescue mode

Once rescue mode is active, you have the same public IP addresses as your instance. You can access the instance as root user via:

• SSH using the temporary password shown in the rescue banner
• Web console via the browser
• VNC console if available

The rescue system’s hostname is <prefix>-<instance name> and its login message can carry your provider’s own text. Providers set both under Admin > System > Settings, and the change applies the next time an instance enters rescue mode.

The temporary password is only shown in the rescue banner and in the API response while rescue mode is active. After you exit rescue mode, the password cannot be retrieved.

Common rescue tasks

List attached disks

lsblk

This shows your instance’s disks with their device names (usually /dev/vda for VirtIO or /dev/sda for SATA) and any partitions.

Mount a disk partition

mount /dev/vda1 /mnt

Replace /dev/vda1 with the correct partition from lsblk. For SATA disks, use /dev/sda1 instead.

Use chroot to access the operating system

Once you have mounted the disk:

arch-chroot /mnt

This gives you a shell inside the instance’s file system, allowing you to run system tools and commands as if the instance had booted normally. Exit with exit or Ctrl+D.

Check and repair the file system

fsck /dev/vda1

Use this to find and repair corrupted file system data. Always unmount the partition first if possible.

Reset a Linux user password

Mount the disk and chroot into it (see above), then use:

chroot /mnt passwd

Then enter the new password for the root user (or passwd username for another user).

Repair Windows NTFS file systems

ntfsfix /dev/vda1

This utility can repair minor NTFS corruption and is often a first step before Windows can boot.

Reset a Windows password

chntpw -u Administrator /mnt/Windows/System32/config/SAM

This allows you to reset the password for the Administrator or any other Windows user. The SAM file must be on a mounted NTFS partition at /mnt.

Repair the GRUB boot loader

For Linux instances with GRUB boot problems, mount the disk and chroot into it, then:

arch-chroot /mnt grub-install /dev/vda
update-grub

This reinstalls GRUB to the boot sector and updates the boot configuration.

Exit rescue mode

To return your instance to normal boot:

1. Open the instance page in the user panel.
2. Click the Exit rescue mode button in the Rescue mode banner.
3. The instance will power cycle and boot into its own disk.

After you exit rescue mode:

• The temporary password is discarded and can never be retrieved.
• Your instance resumes normal operation using its original configuration.

Rescue mode limits

• KVM only: Rescue mode is available for KVM instances. Proxmox instances are not supported.
• Instance state: You cannot enter rescue mode while the instance is suspended, undergoing a live migration, or running an active task.
• Blocked operations: While in rescue mode, you cannot destroy, reinstall, or reset the password of the instance until you exit rescue mode first.
• One-time password: The rescue password is only shown immediately after entering rescue mode and in the API response. After you exit rescue mode, the password is permanently discarded.

See also

• Manage an instance for other power actions and settings.
• Create an instance if you need to deploy a new instance instead.