How to Install and Configure HAProxy on CentOS 8 / RHEL 8

HAProxy stands for High Availability Proxy, it is a free and open source load balancer tool which allow to balance the incoming traffic (TCP and HTTP based) by distributing across the backend servers using different criteria.

In other words, we can say HAProxy is used to provide fault tolerance and high availability in case when one node is getting too many concurrent requests. It is used by most famous web sites like GitHub, Stack Overflow and Tumbler.

In this article, we will discuss how to install and configure HAProxy on CentOS 8/RHEL 8 for Nginx Web Servers. Following are the details for my haproxy lab setup,

HAProxy Server – 192.168.1.10 (haproxy-centos8)
Nginx Server 1 – 192.168.1.11 (nginx-node01)
Nginx Server 2 – 192.168.1.12 (nginx-node01)

Let’s jump into installation and configuration steps of HAProxy on CentOS 8 / RHEL 8

Step:1) Update /etc/hosts file of your HAProxy Server

192.168.1.10    haproxy-centos8
192.168.1.11    nginx-node01
192.168.1.12    nginx-node02

After updating the hosts file, make sure you are able to ping Nginx nodes,

Step:2) Install and Configure HAProxy on CentOS 8 / RHEL 8

Haproxy package is available in the default package repositories of CentOS 8 and RHEL 8, so it can be easily installed with dnf command. But it is recommended update your system before installing haproxy. So execute the following command,

root@haproxy-centos8 ~]# dnf update -y
root@haproxy-centos8 ~]# reboot

Now use following dnf command to install haproxy,

[root@haproxy-centos8 ~]# dnf install haproxy

Once the haproxy is installed successfully, configure it by editing its configuration file “/etc/haproxy/haproxy.cfg“.

Before editing the file, it is always a good practice to make a backup of original file , run the following command,

[root@haproxy-centos8 ~]# cd /etc/haproxy/
[root@haproxy-centos8 haproxy]# cp haproxy.cfg haproxy.cfg-org
[root@haproxy-centos8 haproxy]#

In configuration file, we will modify two sections frontend and backend. In Frontend section we define the haproxy IP and its port, stats uri and backend name. In Backend section we define the type of load balance algorithm we will use like round robin & least connection etc and backend server’s name, IPs and port, example is demonstrated below,

[root@haproxy-centos8 haproxy]# vi haproxy.cfg
…………………
frontend http_balancer
    bind 192.168.1.10:80
    option http-server-close
    option forwardfor
    stats uri /haproxy?stats

#    acl url_static       path_beg       -i /static /images /javascript /stylesheets
#    acl url_static       path_end       -i .jpg .gif .png .css .js
#    use_backend static          if url_static
    default_backend     nginx_webservers

backend nginx_webservers
    mode        http
    balance     roundrobin
    option httpchk HEAD / HTTP/1.1\r\nHost:\ localhost    
    server  nginx-node01  192.168.1.11:80  check
    server  nginx-node02  192.168.1.12:80  check
 ………………………………

Save and exit the file

Configure rsyslog so that it stores all HAProxy statistics, edit rsyslog config file “/etc/rsyslog.conf” and uncomment line 19 and 20,

[root@haproxy-centos8 ~]# vi /etc/rsyslog.conf
……
module(load="imudp")
input(type="imudp" port="514")
……

Save and exit the file.

Now create haproxy.conf file for rsyslog , paste the following lines,

[root@haproxy-centos8 ~]# vi /etc/rsyslog.d/haproxy.conf
local2.=info     /var/log/haproxy-access.log
local2.notice    /var/log/haproxy-info.log

save and exit the file

Restart and enable rsyslog service so that it will be available across the reboots

[root@haproxy-centos8 ~]# systemctl restart rsyslog
[root@haproxy-centos8 ~]# systemctl enable rsyslog

Now finally start haproxy but before starting haproxy service, set the following selinux rule,

[root@haproxy-centos8 ~]# setsebool -P haproxy_connect_any 1

Start & enable haproxy using beneath systemctl commands

[root@haproxy-centos8 ~]# systemctl start haproxy
[root@haproxy-centos8 ~]# systemctl enable haproxy

Allow the haproxy port (in our case it is 80) in OS firewall, execute the beneath firewall-cmd command,

[root@haproxy-centos8 ~]# firewall-cmd --permanent --add-port=80/tcp
[root@haproxy-centos8 ~]# firewall-cmd --reload

Till now HAProxy installation and configuration part is completed let’s move to Nginx nodes,

Step:3) Install NGINX and start its service

# dnf install nginx -y
# systemctl start nginx
# systemctl enable nginx

Let’s modify the index.html file of respective node,

For nginx-node01

[root@nginx-node01 ~]# cd /usr/share/nginx/html
[root@nginx-node01 html]# echo "Nginx Node01 - Welcome to First Nginx Web Server" > index.html

For nginx-node02

[root@nginx-node02 ~]# cd /usr/share/nginx/html
[root@nginx-node02 html]# echo "Nginx Node02 - Welcome to 2nd Nginx Web Server" > index.html

Allow Nginx port (80) in the OS firewall in both nodes using below command,

# firewall-cmd --permanent --add-service=http
# firewall-cmd --reload

Step:4) Test Whether Your HAProxy is working properly or not

Login to haproxy server and run the curl command couple of times to see whether traffic is distributed in round-robin way

[root@haproxy-centos8 ~]# curl 192.168.1.10

Perfect, this confirms haproxy is working properly as it is distributing traffic between two nodes,

Let’s verify from the Web browser too,

Above confirms that HAProxy has configured successfully on CentOS 8 and RHEL 8.

You can view the status of your haproxy via web browser, type url : http://<HA-Proxy-IP>/haproxy?stats

In our case URL will be http://192.168.1.10/haproxy?stats

We can also view HAProxy statistics from log file ( /var/log/haproxy-access.log).

That’s all from this tutorial, I hope these steps help you to setup HAProxy on CentOS 8 and RHEL 8 smoothly. Please do share your valuable feedback and comments.

Also Read: How to Setup Django Python Framework on CentOS 8