Monitoring Data - Big Data Made Easy: A Working Guide to the Complete Hadoop Toolset - page 214

It is important to reiterate Nagios's alerting capability because not only can you determine that there is a problem
but you can also create an automatic alert and raise awareness of the issue. So, the combination of Ganglia and Nagios
for system monitoring is a highly functional duo. Nagios alerts you to a problem while Ganglia provides a color-coded
graph to display the problem. A typical problem might be that Hadoop has run out of memory or disk space. Now, you
can be alerted to the problem and assess the potential solutions.
Installing Nagios
To demonstrate how to install Nagios, I place the Nagios server on the name node, hc1nn, and the Nagios clients on all
of the servers so that the whole cluster will be monitored and the Nagios map (more on this in a moment) will contain
the entire cluster. So, on hc1nn, I install the following servers via the Linux yum command as root:
[root@hc1nn ~]# yum install nagios
[root@hc1nn ~]# yum install nagios-plugins-all
[root@hc1nn ~]# yum install nagios-plugins-nrpe
[root@hc1nn ~]# yum install nrpe
In your installation, you will also need the php and httpd components, which you probably already have
installed. When you execute the following commands, you will receive a “Nothing to do” message if the components
are already installed:
[root@hc1nn ~]# yum install php
[root@hc1nn ~]# yum install httpd
I use the chkconfig command to configure these servers (httpd, nrpe, and Nagios) to start when the server hc1nn
is rebooted. The Nagios server uses the nrpe components on each machine, which in turn use the Nagios plug-ins to
monitor different features, like hosts, devices, and services:
[root@hc1nn ~]# chkconfig httpd on
[root@hc1nn ~]# chkconfig nagios on
[root@hc1nn ~]# chkconfig nrpe on
Now, I install the the following Nagios components on the data nodes (hc1r1m1, hc1r1m2, hc1r1m3) by using the
Linux yum command as root:
[root@hc1nn ~]# yum install nagios
[root@hc1nn ~]# yum install nagios-plugins-all
[root@hc1nn ~]# yum install nrpe
[root@hc1nn ~]# chkconfig nrpe on
On the name node, I set the password for the Nagios administration account, nagiosadmin. I will use this account
and password combination to access the Nagios web browser:
[root@hc1nn ~]# htpasswd -c /etc/nagios/passwd nagiosadmin