--- title: "Floating IPs" description: "Portable public IPs owned by your organization: acquire, assign, set rDNS, release, and configure the address inside the guest OS when needed." category: "Security" language: en source: https://cubepath.com/docs/products/security/floating-ips updated: 2026-11-17 --- # Floating IPs A Floating IP is a portable public IP address owned by your organization. You can assign it to one machine, then move it to another at any time without changing the address. Manage them on the **Floating IPs** page in [my.cubepath.com](https://my.cubepath.com). ## What it's for Because the IP belongs to your organization and not to a single machine, it keeps the same address across rebuilds and migrations. That makes it the right tool for: - **Stable DNS records**, so a rebuild doesn't mean editing DNS and waiting out a TTL. - **Failover**, moving the address to a standby machine in seconds. - **Migrations**, cutting traffic over to a new server without touching anything that points at the old one. - **Whitelisted addresses**, where a partner has registered your IP and changing it means a support ticket on their side. > Note: A Floating IP only attaches to machines in the **same location**. An IP acquired in one region can't be assigned to a server in another. ## Acquire an IP Click **Acquire IP**, choose the **IP Type** (IPv4) and the **Location**, then confirm. The new address appears under the **Single IPs** tab. ![Acquire Floating IP with price, IP type and location](https://cdn.cubepath.com/img/docs/products/2681654c-f28a-49ad-b47f-4315d80c67cb.png) Larger allocations are grouped as **Subnets**, where you can expand a range to see and manage every individual IP it contains. ## Assign and unassign The **Assigned to** column shows the VM or dedicated server an IP is currently attached to, or **Unassigned** if it's free. To attach an IP, open the machine's **Network** tab and click **Add IP** in the **Floating IPs** section; only free IPs in the machine's location are listed. To detach one, use the **Unassign IP** button next to it on the same tab; the IP stays in your organization. To move an IP to another machine, unassign it from the first one and add it to the second. > Note: Restart the server after assigning or unassigning an IP so the change takes effect on the machine. ## Configure the IP inside the guest OS Most of the time the machine picks the address up on reboot and there's nothing to do. When it doesn't (a static-networking image, or an extra address on a machine that already has one), add it by hand. First check what the OS actually sees: ```bash ip -br addr ``` Then add the address persistently, using whatever your distribution uses: - **Ubuntu (netplan)**: add the address to the interface block in `/etc/netplan/*.yaml`, then test with `sudo netplan try`. It auto-reverts after 120 seconds if you lose SSH, which turns a mistake into a two-minute wait instead of a console session. - **Debian (interfaces)**: add the address to `/etc/network/interfaces`, then `sudo systemctl restart networking`. - **RHEL / CentOS / Rocky**: use `nmcli`, or add it to the interface's configuration under `/etc/sysconfig/network-scripts/` and restart networking. Additional addresses are configured with a **/32** mask (IPv4) or **/128** (IPv6): they're routed to your machine individually, not as part of a local subnet. The default gateway stays the one your primary address already uses; don't add a second one. > Warning: Editing network configuration over SSH is how people lock themselves out. Have the [browser console](/docs/products/deployment/deploy-a-virtual-machine#access-your-machine) open in another tab before you restart networking; it doesn't depend on the configuration you're about to change. Verify with `ip -br addr` again, and from outside with a ping or a connection to a service bound to the new address. ## Reverse DNS (rDNS) For an assigned IP, use the edit control next to **Reverse DNS** to set its PTR record (for example `mail.example.com`). Leave it empty to remove the record. This matters most for **outbound mail**: many receiving servers reject or spam-folder messages from an IP whose reverse lookup doesn't resolve, or doesn't match the sending hostname. It's also worth setting on a [NAT Gateway](/docs/products/deployment/nat-network)'s egress address, so a partner can tell which of your addresses is calling their API. ## Release an IP Open the **Manage** menu of a secondary IP and choose **Release IP** (a machine's primary IP can't be released). The address returns to the available pool and **billing stops**. > Warning: A released address goes back to the general pool and will eventually be assigned to someone else. Anything still pointing at it (DNS records, partner whitelists, firewall rules) should be updated first. When you delete a VPS, its floating IPs are released unless you tick **Keep floating IPs in my organization** in the confirmation dialog. Keeping them means they keep billing but stay yours; releasing them stops the charge but gives the addresses up. > Tip: Need DDoS protection? Choose **Configure Protection** in the **Manage** menu of an IP or subnet to set up its protection in [DDoS Mitigation](/docs/products/security/ddos-mitigation).