How to Install and Configure Keepalived for High Availability

In this blog post, I’ll guide you through setting up Keepalived for high availability on a multi-node system. Keepalived uses the VRRP (Virtual Router Redundancy Protocol) to ensure a seamless failover between a primary and backup node, allowing you to manage a virtual IP address that switches automatically if the primary node becomes unavailable.

We’ll cover the following:

1. Installing Keepalived on both the primary and backup nodes.
2. Configuring Keepalived for automatic failover and failback.
3. Verifying failover behavior.

Step 1: Install Keepalived on All Nodes

First, install Keepalived on both the primary and backup nodes using apt:

apt -y install keepalived

Repeat this on both nodes to ensure they are running the Keepalived service.

Step 2: Configure Keepalived on the Primary Node

On the primary node, you’ll configure Keepalived to act as the MASTER. Here’s the configuration you need to add to the /etc/keepalived/keepalived.conf file:

vrrp_instance VRRP1 {
    state MASTER
    interface ens192
    virtual_router_id 101
    priority 200
    advert_int 1
    virtual_ipaddress {
        10.34.104.101/24
    }
}

Breakdown:

• state MASTER: This node is the primary node that holds the virtual IP.
• interface ens192: Network interface you are using.
• virtual_router_id 101: Unique ID for this VRRP instance.
• priority 200: Higher priority ensures that this node is preferred as the master.
• virtual_ipaddress: The virtual IP address assigned for failover.

Once you’ve made these changes, restart the Keepalived service:

systemctl restart keepalived

Verify that the virtual IP is assigned to the interface:

ip address show ens192

You should see the virtual IP 10.34.104.101 assigned to ens192.

Step 3: Configure Keepalived on the Backup Node

On the backup node, configure Keepalived to act as the BACKUP. Modify /etc/keepalived/keepalived.conf as follows:

vrrp_instance VRRP1 {
    state BACKUP
    interface ens192
    virtual_router_id 101
    priority 100
    advert_int 1
    virtual_ipaddress {
        10.34.104.101/24
    }
}

Breakdown:

• state BACKUP: This node takes over if the master becomes unavailable.
• priority 100: Lower priority ensures this node only takes over when the master is down.

Restart Keepalived on the backup node:

systemctl restart keepalived

Step 4: Verify Failover and Failback

Simulating Failover

To test the failover, bring down the primary node’s interface:

ip link set down ens192

On the backup node, check the network interface:

ip address show ens192

You should see the virtual IP 10.34.104.101 assigned to the backup node’s ens192 interface, confirming the failover.

Simulating Failback

Now, bring the primary node’s interface back up:

ip link set up ens192

On the primary node, verify that the virtual IP has been reassigned:

ip address show ens192

The virtual IP should be back on the primary node, indicating that failback is working correctly.