---
title: "Recovery Mode"
---

# Accessing a Node in Recovery Mode

Recovery Mode boots the node from a rescue disk image instead of its normal disk. This gives you a working shell even when the regular OS will not boot — for example, after a botched config change, a full root disk, or a forgotten root password. This guide walks through enabling Recovery Mode, mounting the disk, performing common repairs, and exiting cleanly.

## Common Issues with Recovery Mode

| Issue                                       | What to do                                                                                 |
| ------------------------------------------- | ------------------------------------------------------------------------------------------ |
| **Enable Recovery Mode** asks for power off | The node must be Powered Off before the rescue disk can be attached. Power off and retry.  |
| Recovery Mode state remains "in progress"   | Refresh the page after a couple of minutes. If the state does not settle, contact support. |
| Cannot change password — prompt is not root | Run `sudo -i` and retry the `passwd` command.                                              |

## When to Use Recovery Mode

- The node will not boot from its primary disk.
- The disk is full and SSH/console is unresponsive.
- You have forgotten the root password and need to reset it.
- A bad firewall or network change has cut you off and the change must be undone from outside the running OS.

## Step 1: Enable Recovery Mode

Open **Manage Nodes**, locate the node, and click **Actions**.

### Power Off the Node First

Recovery Mode boots from a different disk, so the node must be powered off before it can be enabled. Click **Power Off** under Actions.

### Enable Recovery Mode

Once the node is Powered Off, open **Actions** again and select **Enable Recovery Mode**.

Click **OK** to confirm.

The node state will move to **Recovery Mode in Progress**.

When the state settles to **Recovery Mode Enabled**, the rescue disk is attached.

## Step 2: Start the Node into Recovery

Click **Start** to boot the node from the rescue disk.

When the node is Running, click **Access Console**.

## Step 3: Boot the Rescue Environment

On the boot menu, navigate to **Troubleshooting** with the arrow keys and press **Enter**.

Select the rescue option, such as **Rescue a Rocky Linux system**, and press **Enter** to boot the rescue environment. The exact label can vary with the rescue image currently attached by the platform.

:::note
If the rescue disk has been enabled but the boot menu is left untouched, the installer may offer to start a new installation. **Do not** continue — quit the installation and return to the rescue flow.
:::

After quitting, the rescue environment will continue booting.

## Step 4: Mount Your Disk

When the rescue prompt appears, choose option **1** to mount your disk. Confirm by entering `1` on the next screen.

Press **Enter** to get the command line in the rescue environment. Your node's disk is now mounted at **`/mnt/sysroot`**.

## Common Recovery Tasks

### Free Up Disk Space

Use `df -Th` to verify the disk is mounted and to check usage:

```bash
df -Th
```

The node's primary disk (`vda1`) appears under `/mnt/sysroot`.

Browse `/mnt/sysroot` and delete the files you no longer need. Log files under `/mnt/sysroot/var/log/` and backup directories are usually the largest consumers.

### Reset the Root Password

To change the root password, change root into the mounted disk and run `passwd`:

```bash
chroot /mnt/sysroot
passwd
```

:::note
Verify the prompt shows `root@localhost` before changing the password. If it does not, run `sudo -i` first — the root password must be changed as root.
:::

## Step 5: Exit and Reboot Normally

When you are done, unmount and exit the rescue environment:

```bash
exit
```

Then in MyAccount:

1. **Power off** the node.

2. Open **Actions** and click **Disable Recovery Mode**.

3. **Start** the node again — it will now boot from its primary disk.

4. Confirm the node returns to **Running** state.

## Related Resources

- [Security Troubleshooting](./security.md)
- [Disk Space Issues](./disk-space.md)
- [Node Not Accessible](./node-not-accessible.md)

---