Moving between nodes

A VPS moves to another node from its page: Administration › Move to another node. The panel runs the checks first (room, addresses, private networks, mode) and shows what will happen; Move queues the move. Moves run one at a time.

Live and cold

• Live: the VPS stays on and doesn't restart; it pauses for an instant only at the final switch. It needs a VPS without additional disks, started by the panel with live moves enabled (see below), and nodes with the same CPU. If live doesn't work, the panel moves it cold on its own (unless you asked for live only).
• Cold: on ZFS the panel copies everything while the VPS runs, starting from a temporary snapshot, then stops it, copies only the changes (usually seconds) and starts it on the new node. The downtime is the shutdown, the last copy and the system boot. Additional disks and snapshots are copied with the VPS.

Data passes through the panel, encrypted: nodes don't need to reach each other. If something goes wrong before the final switch, the VPS stays as it was on the old node.

Live moves are enabled in Settings (on by default): they apply from the next start made by the panel (creation, start, restart, reinstall), for VPSs without additional disks. With the first additional disk the VPS restarts once and can no longer move live. On the VPS page, the Live row shows whether it's ready.

Addresses

• The IPv4 stays if its block is also attached to the target node, or if it's a provider IP that can be moved to the new node's server (1–3 more minutes without network).
• Otherwise tick Change the addresses if needed: the VPS gets free IPv4, IPv6 and additional IPs on the new node and the panel changes the network inside the VPS with the agent. Without IPv6 on the new node, the VPS has none. VPSs with a hand-made network setup (or Windows) must be updated by hand.
• Private networks stay the same: the panel sets up the tunnels on the new node if needed.
• Backups stay where they are; the next incremental backup starts again with a full copy.

Drain node

In Nodes, Drain node puts the node into maintenance and moves all its VPSs: for each one the panel picks a node in the same city, then the emptiest, and shows you the list with mode and problems first. VPSs that can't be moved stay (with the reason). Stop draining cancels the moves not yet started; the node stays in maintenance.

Rebalancing

Below the nodes, Rebalancing compares assigned RAM and disk (the higher of the two) and suggests moves from the fullest node to the emptiest, as long as the difference is above 20%, without changing addresses and preferring VPSs that can move live. Apply queues one move, Apply all all of them.