Recovery mode

Recovery mode starts a node from a small rescue disk instead of its own. Reach for it when:

• You have lost or forgotten the node’s password.
• The node’s disk is so full that SSH and the console have stopped responding.
• The node will not boot from its own disk at all.
• A firewall or network change locked you out, and undoing it means working from outside the system that is still running with that change in place.

What turning on recovery mode does

Enable Recovery Mode attaches a rescue disk that matches the node’s operating system and region, then boots the node from that disk instead of its own. Because the fix runs from a separate disk, it works even when the node’s own disk or boot setup is what is broken. Recovery mode is tracked separately from the node’s main status: the node keeps showing its regular state (for example Powered off) the whole time it is on. Turning recovery mode off detaches the rescue disk again, and the node goes back to booting from its own disk the next time it starts.

Enable Recovery Mode appears once the node is Powered off, for Linux nodes: CentOS, Debian, Ubuntu, openSUSE, CloudLinux, Red Hat, AlmaLinux, and Rocky Linux. Windows, FortiGate, and spot nodes do not offer it, and you cannot turn it on again while it is already on or still turning on. See Spot nodes and FortiGate firewall nodes for what those are.

Getting in and fixing the node

Reach the node through Access Console once recovery mode has finished starting up. Access Console is the in-browser view of a node’s screen, like VNC, used when SSH or RDP will not connect. It stays disabled for the short time it takes to boot from the rescue disk; if you try an action too soon, you will see a message asking you to retry in a few seconds. Wait and try again.

From the rescue disk you can reset a password you are locked out of, free up space on a full disk, or repair whatever is stopping the node from starting on its own. Enable recovery mode has the steps, including the password-reset procedure.