For Linux/Unix platforms only:

WARNING: FusionInventory Agent would better be installed as systemd service and
this is the recommanded way of doing.

FusionInventory agent is expected to run its tasks on a regular basis and is better
installed as service. Its memory and load footprints are low enough for modern desktop computers.

To know better about the FusionInventory agent execution modes, you should read
the online Agent usage page:
http://fusioninventory.org/documentation/agent/usage.html

You should use cron if you don't want a fusioninventory-agent process to be running
all the time and if you don't need to trigger it remotely (from the server
or manually, locally contacting the `http://localhost:62354/` page).

So if you prefer to run FusionInventory Agent as scheduled cron job, you may need
to do the following as root:
1. stop and disable the fusioninventory-agent service (if start/enabled):
   systemctl stop fusioninventory-agent
   systemctl disable fusioninventory-agent

2. create a /etc/fusioninventory/conf.d/better-for-cron.cfg file with the following
   lines:
   # Don't force agent to run inventory at each run, but only
   # after server defined timeout has been reached
   lazy = 1

3. as example, add a root cron job like this:
   # Don't remove --wait option (delay in seconds) so the agent wait a random
   # time before really run its work
   00 */4 * * * /usr/bin/fusioninventory-agent --wait=1800 >/var/lib/fusioninventory-agent/cron.log 2>&1

For large companies, be aware you can lead to concurrent execution issues
if you don't enable agent `lazy` option with cron scheduling and even more if you forget to
use `--wait` option. To avoid such missconfiguration, you would prefer leave the agent
started as systemd service and define a peaceful delay on server. You may adjust scheduled
time and wait command line only option to your needs.

As a reminder, when launched without the daemonize configuration directive
as usually done from cron, the agent is a short-lived process, without
the ability to listen on the network for out-of-schedule activation requests
from the server.
===

Review our mandatory 'schedule agent with cron' page:
http://fusioninventory.org/documentation/agent/cron.html

For Linux/Unix platforms packaging only:

FusionInventory Agent should be installed as systemd service by default.

By default, service will fail until agent configuration includes a server target
configuration or a local target configuration.

The default configuration should use the new "include" directive so any *.cfg file
from /etc/fusioninventory/conf.d folder is read to update default options values.

Then the best way to configure a target will be to create a dedicated .cfg file
in /etc/fusioninventory/conf.d folder. As example, users would have to create a
simple glpi.cfg file set with the following:

# GLPI server
server = http://glpi-server/plugins/fusioninventory

Or for a local inventory:

# Local inventory folder
local = /var/lib/fusioninventory-agent

If "include" directive is enabled at the end of default configuration, any
new definition set in included file(s) will just override defaults, so as example
anyone can just enable debug level 1 option and set tag adding the following to a
cfg file or adding it in a newer cfg file like this:

debug = 1
tag = hosting

Be aware, files in /etc/fusioninventory/conf.d folder would be read in alphabetic
order and last value for a specific option will be used for the option.

===

Default logger configuration is to log on stderr. Packagers should prefer the
agent to use system logging feature while installed as service and setting the
following in installed default configuration file:
logger = syslog
logfacility = LOG_DAEMON

===

During packaging, it is also advised to override lib/FusionInventory/Agent/Version.pm
module so agent version matches the package version. Packagers should also add
their own useful comments in the $COMMENTS perl array ref to help maintainers
to identify the package building source. As example, the module can be set to:
package FusionInventory::Agent::Version;

use strict;
use warnings;

our $VERSION = "2.4.2-1";
our $PROVIDER = "FusionInventory";
our $COMMENTS = [ "My distro package built on some computer", "Built on that date" ];

1;

===

During packaging, it is also advised to override lib/setup.pm module so datadir,
libdir and vardir perl variables match the system preferences for these folders.
The default setup.pm is defined to compute most well-known case with a very light
overhead.

===

Review our mandatory Linux packaging page:
http://fusioninventory.org/documentation/packaging/linux.html

# FusionInventory Agent

[![Travis Build Status](https://travis-ci.org/fusioninventory/fusioninventory-agent.svg?branch=develop)](https://travis-ci.org/fusioninventory/fusioninventory-agent)
[![Appveyor Build status](https://ci.appveyor.com/api/projects/status/f2oh6p3qnr2bck1b?svg=true)](https://ci.appveyor.com/project/fusioninventory/fusioninventory-agent)
[![CircleCI Build status](https://circleci.com/gh/fusioninventory/fusioninventory-agent.svg?style=svg)](https://circleci.com/gh/fusioninventory/fusioninventory-agent)

## Summary

The FusionInventory agent is a generic management agent. It can perform a
certain number of tasks, according to its own execution plan, or on behalf of a
GLPI server with fusioninventory plugin, acting as a control point.

## Description

See [FusionInventory solution overview](http://fusioninventory.org/overview/)

## Dependencies

### Core

Minimum perl version: 5.8

Mandatory Perl modules:

* File::Which
* LWP::UserAgent
* Net::IP
* Text::Template
* UNIVERSAL::require
* XML::TreePP

Optional Perl modules:

* Compress::Zlib, for message compression
* HTTP::Daemon, for web interface
* IO::Socket::SSL, for HTTPS support
* LWP::Protocol::https, for HTTPS support
* Proc::Daemon, for daemon mode (Unix only)
* Proc::PID::File, for daemon mode (Unix only)

### Inventory task

Optional Perl modules:

* Net::CUPS, for printers detection
* Parse::EDID, for EDID data parsing
* DateTime, for reliable timezone name extraction

Optional programs:

* dmidecode, for DMI data retrieval
* lspci, for PCI bus scanning
* hdparm, for additional disk drive info retrieval
* monitor-get-edid-using-vbe, monitor-get-edid or get-edid, for EDID data access
* ssh-keyscan, for host SSH public key retrieval

### Network discovery tasks

Mandatory Perl modules:

* Thread::Queue

Optional Perl modules:

* Net::NBName, for NetBios method support
* Net::SNMP, for SNMP method support

Optional programs:

* arp, for arp table lookup method support

### Network inventory tasks

Mandatory Perl modules:

* Net::SNMP
* Thread::Queue

Optional Perl modules:

* Crypt::DES, for SNMPv3 support

### Wake on LAN task

Optional Perl modules:

* Net::Write::Layer2, for ethernet method support

### Deploy task

Mandatory Perl modules:

* Archive::Extract
* Digest::SHA
* File::Copy::Recursive
* JSON::PP
* URI::Escape

Mandatory Perl modules for P2P Support:
* Net::Ping
* Parallel::ForkManager

## Related contribs

See [CONTRIB](CONTRIB.md) to find references to FusionInventory Agent related scritps/files

## Contacts

Project websites:

* main site: <http://www.fusioninventory.org>
* Github org: <http://github.com/fusioninventory/>

Project mailing lists:

* <http://lists.alioth.debian.org/mailman/listinfo/fusioninventory-user>
* <http://lists.alioth.debian.org/mailman/listinfo/fusioninventory-devel>

Project IRC channel:

* #FusionInventory on FreeNode IRC Network

Please report any issues on Github.

## Active authors

* Guillaume Rousse <guillomovitch@gmail.com>
* Guillaume Bougard <gbougard@teclib.com>

Copyright 2006-2010 [OCS Inventory contributors](https://www.ocsinventory-ng.org/)

Copyright 2010-2019 [FusionInventory Team](http://fusioninventory.org)

Copyright 2011-2019 [Teclib Editions](http://www.teclib-edition.com/)

## License

This software is licensed under the terms of GPLv2+, see LICENSE file for
details.

## Additional pieces of software

The fusioninventory-injector script:

* author: Pascal Danek
* copyright: 2005 Pascal Danek

FusionInventory::Agent::Task::Inventory::Input::Virtualization::Vmsystem
contains code from imvirt:

* url: <http://micky.ibh.net/~liske/imvirt.html>
* author: Thomas Liske <liske@ibh.de>
* copyright: 2008 IBH IT-Service GmbH <http://www.ibh.de/>
* License: GPLv2+