For mission-critical applications, relying on a single server is a risk. Tilaa offers Failover IP solutions that allow traffic to automatically switch to a standby server if the primary server fails.
[Image of firewall allowing traffic]1. Choose your Architecture
We offer two types of redundancy. Pricing can be found on our website.
Option A: Single Datacenter Failover
Scenario: Two servers in the same location (e.g., Amsterdam).
Setup: A new Floating IP is assigned to your account. You can route this IP to Server A or Server B via the API or Keepalived.
Option B: Cross-Datacenter Failover (Twin-DC)
Scenario: One server in Haarlem, one in Amsterdam.
Setup: This requires a more advanced network configuration.
- Tilaa assigns a specific new network interface to your servers to handle this traffic.
- Action Required: You must reboot your servers via the Tilaa Dashboard so the new network card is detected by the OS.
2. OS Configuration (Critical: Asymmetric Routing)
When using Failover IPs, traffic might enter via one interface but leave via another. By default, Linux security features block this behavior.
You must disable
rp_filter in your kernel settings. If you skip this, the server will silently drop the failover traffic.
1. Edit the sysctl configuration:
nano /etc/sysctl.conf2. Add or modify these lines:
net.ipv4.conf.default.rp_filter = 0
net.ipv4.conf.all.rp_filter = 03. Apply changes immediately:
sysctl -p3. Failover Software (VRRP / Keepalived)
Having the IP capability is step 1. You need software on your servers to actually monitor health and switch the IP.
We recommend using Keepalived (based on the VRRP protocol). It sends "heartbeats" between your servers. If the primary server stops responding, the secondary server automatically claims the IP address.
Configuring High Availability clusters is an advanced task.
- Do it yourself: Follow our Guide to configuring Keepalived.
- Hire an Expert: Our support engineers can set this up for you as a paid consultancy service. Please contact Sales for a quote.