Thread: Nagios Kit

Overview

The Nagios kit installs the Nagios monitoring package on your cluster. From the official website:

Nagios® is an Open Source host, service and network monitoring program.

For more information about Nagios, visit About Nagios or the official documentation.

Installation

The kit should be installed as usual.

Accessing Nagios Status Page

From the monitoring node, the Nagios control panel is available at http://localhost/nagios/. The default login and password are admin.

Kit Contents


The kit contains the following packages from EPEL:

• Nagios
• Nagios plugins
• Nagios Remote Plugin Executor (NRPE)

Two additional packages, kusu-nagios-extras and kusu-nrpe-extras, are also included. These packages contain tools and scripts which configure Nagios to run out of the box when it is first installed. The kit package (kit-nagios) contains plugins for the addhost and ngedit tools, used to reconfigure Nagios whenever the cluster layout is modified.


There are two components this kit provides, component-installer-nagios and component-compute-nagios. They are intended to be installed on monitoring and monitored nodes, respectively. component-installer-nagios installs Nagios and the plugins, while component-compute-nagios installs NRPE and the plugins.

Customizing Configuration

Nagios configuration is generated by the nagiosconfig.py program. Two directories in /etc/nagios are accessed by this program: commands and nodes.

Nodes

The nodes subdirectory contains config files with .cfg extensions as well as template files with .tpl extensions. The template files are used to generate actual configuration files for nodes and nodegroups. Kusu nodegroups map to Nagios Host Groups and Kusu nodes map to Nagios Hosts.


The template files are used to generate the actual configuration files for each node/nodegroup. The templates currently have a few variables which are replaced with valid data in the configuration file.

• $nodegroup_name is set to the name of the Kusu nodegroup
• $nodegroup_description is also set to the Kusu nodegroup name
• $nodegroup_nodes is a listing of monitored nodes in the nodegroup
• $node_name is the hostname of the node
• $node_alias is also the hostname of the node
• $node_ip is also the hostname of the node; this is used to access NRPE on remote nodes so should resolve properly
• $service_definitions is replaced with services to check for this node generated from templates in the commands subdirectory

Note the node's IP is not used to access the node because it could be non-trivial to discover over which network to perform monitoring. The hostname should be resolvable by any node in the cluster.

Commands

The commands subdirectory contains config files with .cfg extensions as well as template files with .tpl extensions. The config files are command definitions for Nagios. The template files are used to generate the $service_definitions section of node configuration files.


The files in the commands directory can have one of three prefixes:

• check_
• installer_
• compute_

Templates prefixed with check_ are used when generating configuration files for any node. Templates with an installer_ prefix are only used when generating configuration files for nodes belonging to installer-type nodegroups as specified in the Kusu database. Similarly, templates prefixed with compute_ are used when generating configuration files for nodes in compute-type nodegroups.


To add custom commands, make sure to provide a configuration file describing the command as well as a template file to use when generating the $service_definitions section of a node configuration file.

How Is Configuration Determined & Distributed?

The Nagios kit package (kit-nagios) includes plugins for the addhost and ngedit tools. These plugins will generate a meta-configuration file every time there is a relevant to Nagios change in the cluster. The file is called nagios.pickle and is placed in /etc/cfmsync/<ng_names>/etc/nagios for all nodegroups which are assigned component-installer-nagios.

The file is a pickled Python dictionary of nodes Nagios is to monitor. Only nodes having at least one Nagios component associated will be listed.

The plugins call cfmsync -f to distribute the new meta-configuration file to any nodes performing monitoring. A cfmclient plugin on each such node calls nagiosconfig.py to regenerate configuration files and reload the Nagios daemon.

Links

• Nagios homepage: http://www.nagios.org
• EPEL homepage: http://fedoraproject.org/wiki/EPEL

Where are the kits for i386? I have Kusu 1.1 and centos 5.4 installed but none of the kits installed. The two nodes PXE boot ok as well.

smtp is your mail service on port 110, set up outlook and when it asks who and where tell it what you told me