Master Keepalived: Step-by-Step HA Setup for Nginx with VIP Failover

Keepalived Overview

Keepalived is a layer‑3/4/5 switching software that monitors web servers. When a server fails, it removes the faulty node from the pool and automatically restores it when it recovers, without manual intervention.

Environment Preparation

Two CentOS 6.6 64‑bit servers are used:

Operating system: CentOS 6.6 64‑bit (2 nodes)

Nginx‑Master 10.0.0.60

Nginx‑Backup 10.0.0.61

Virtual IP 10.0.0.62

Both nodes have identical configuration unless otherwise noted.

Install Nginx

Use OneinStack to install Nginx, selecting “y” for Nginx and “n” for other components.

Install Keepalived

On both Nginx‑Master and Nginx‑Backup execute:

cd ~/oneinstack/src

wget http://www.keepalived.org/software/keepalived-1.2.22.tar.gz

tar xzf keepalived-1.2.22.tar.gz

cd keepalived-1.2.22

./configure --prefix=/usr/local/keepalived

make && make install

Configure Keepalived

Create symbolic links and enable the service:

ln -s /usr/local/keepalived/etc/keepalived /etc/keepalived

ln -s /usr/local/keepalived/etc/rc.d/init.d/keepalived /etc/rc.d/init.d/keepalived

ln -s /usr/local/keepalived/etc/sysconfig/keepalived /etc/sysconfig/keepalived

ln -s /usr/local/keepalived/sbin/keepalived /usr/bin/keepalived

chkconfig keepalived on

Edit /etc/keepalived/keepalived.conf on the master node with sections for global_defs (email, SMTP, router_id), a vrrp_script named chk_nginx that runs /usr/local/keepalived/sbin/check_nginx.sh, and a vrrp_instance VI_1 configured as follows:

state MASTER

interface eth0

virtual_router_id 55

priority 100

advert_int 1

authentication PASS/linuxeye

virtual_ipaddress 10.0.0.62

track_script chk_nginx

On the backup node use a similar configuration but set state BACKUP, priority 50, and add nopreempt to prevent VIP preemption.

Health‑Check Script

Create /usr/local/keepalived/sbin/check_nginx.sh:

#!/bin/bash
if [ "$(ps -ef | grep "nginx: master process"| grep -v grep )" == "" ];then
    /etc/init.d/nginx start
    sleep 5
    if [ "$(ps -ef | grep "nginx: master process"| grep -v grep )" == "" ];then
        /etc/init.d/keepalived stop
    fi
fi

Make it executable:

chmod +x /usr/local/keepalived/sbin/check_nginx.sh

Verification

Start Keepalived on both nodes ( service keepalived start), check the virtual IP with ip addr, then stop the service on one node to observe failover. Restart to confirm the VIP moves back.

Advanced Tip

To avoid packet loss when replacing the master, forward traffic to the backup with iptables before stopping the master:

iptables -F

iptables -t nat -I PREROUTING -i eth0 -j DNAT --to-destination 10.0.0.61

iptables -t nat -I POSTROUTING -o eth0 -j MASQUERADE