path: root/docs-xml/manpages/vfs_smb_traffic_analyzer.8.xml

<?xml version="1.0" encoding="iso-8859-1"?>
<!DOCTYPE refentry PUBLIC "-//Samba-Team//DTD DocBook V4.2-Based Variant V1.0//EN" "http://www.samba.org/samba/DTD/samba-doc">
<refentry id="vfs_smb_traffic_analyzer.8">

<refmeta>
	<refentrytitle>smb_traffic_analyzer</refentrytitle>
	<manvolnum>8</manvolnum>
	<refmiscinfo class="source">Samba</refmiscinfo>
	<refmiscinfo class="manual">System Administration tools</refmiscinfo>
	<refmiscinfo class="version">4.1</refmiscinfo>
</refmeta>


<refnamediv>
	<refname>vfs_smb_traffic_analyzer</refname>
	<refpurpose>log Samba VFS read and write operations through a socket
	to a helper application</refpurpose>
</refnamediv>

<refsynopsisdiv>
	<cmdsynopsis>
		<command>vfs objects = smb_traffic_analyzer</command>
	</cmdsynopsis>
</refsynopsisdiv>

<refsect1>
	<title>DESCRIPTION</title>

	<para>This VFS module is part of the
	<citerefentry><refentrytitle>samba</refentrytitle>
	<manvolnum>7</manvolnum></citerefentry> suite.</para>

	<para>The <command>vfs_smb_traffic_analyzer</command> VFS module logs
	client file operations on a Samba server and sends this data
	over a socket to a helper program (in the following the "Receiver"),
	which feeds a SQL database. More
	information on the helper programs can be obtained from the
	homepage of the project at:
	http://holger123.wordpress.com/smb-traffic-analyzer/
	Since the VFS module depends on a receiver that is doing something with
	the data, it is evolving in it's development. Therefore, the module
	works with different protocol versions, and the receiver has to be able
	to decode the protocol that is used. The protocol version 1 was
	introduced to Samba at September 25, 2008. It was a very simple
	protocol, supporting only a small list of VFS operations, and had
	several drawbacks. The protocol version 2 is a try to solve the
	problems version 1 had while at the same time adding new features.
	With the release of Samba 4.0.0, the module will run protocol version 2
	by default.
	</para>
</refsect1>

<refsect1>
	<title>Protocol version 1 documentation</title>
	<para><command>vfs_smb_traffic_analyzer</command> protocol version 1 is aware
		of the following VFS operations:</para>

	<simplelist>
        <member>write</member>
        <member>pwrite</member>
	<member>read</member>
	<member>pread</member>
	</simplelist>

	<para><command>vfs_smb_traffic_analyzer</command> sends the following data
	in a fixed format separated by a comma through either an internet or a
	unix domain socket:</para>
	<programlisting>
	BYTES|USER|DOMAIN|READ/WRITE|SHARE|FILENAME|TIMESTAMP
	</programlisting>

	<para>Description of the records:

	<itemizedlist>
	<listitem><para><command>BYTES</command> - the length in bytes of the VFS operation</para></listitem>
	<listitem><para><command>USER</command> - the user who initiated the operation</para></listitem>
	<listitem><para><command>DOMAIN</command> - the domain of the user</para></listitem>
	<listitem><para><command>READ/WRITE</command> - either "W" for a write operation or "R" for read</para></listitem>
	<listitem><para><command>SHARE</command> - the name of the share on which the VFS operation occurred</para></listitem>
	<listitem><para><command>FILENAME</command> - the name of the file that was used by the VFS operation</para></listitem>
	<listitem><para><command>TIMESTAMP</command> - a timestamp, formatted as "yyyy-mm-dd hh-mm-ss.ms" indicating when the VFS operation occurred</para></listitem>
	<listitem><para><command>IP</command> - The IP Address (v4 or v6) of the client machine that initiated the VFS operation.</para></listitem>
	</itemizedlist>

	</para>

	<para>This module is stackable.</para>

</refsect1>

<refsect1>
	<title>Drawbacks of protocol version 1</title>
	<para>Several drawbacks have been seen with protocol version 1 over time.</para>
	<itemizedlist>
	<listitem>
		<para>
			<command>Problematic parsing - </command>
			Protocol version 1 uses hyphen and comma to separate blocks of data. Once there is a
			filename with a hyphen, you will run into problems because the receiver decodes the
			data in a wrong way.
		</para>
	</listitem>
	<listitem>
		<para>
			<command>Insecure network transfer - </command>
			Protocol version 1 sends all it's data as plaintext over the network.
		</para>
	</listitem>
	<listitem>
		<para>
			<command>Limited set of supported VFS operations - </command>
			Protocol version 1 supports only four VFS operations.
		</para>
	</listitem>
	<listitem>
		<para>
			<command>No subreleases of the protocol - </command>
			Protocol version 1 is fixed on it's version, making it unable to introduce new
			features or bugfixes through compatible sub-releases.
		</para>
	</listitem>
	</itemizedlist>
</refsect1>
<refsect1>
	<title>Version 2 of the protocol</title>
	<para>Protocol version 2 is an approach to solve the problems introduced with protcol v1.
	From the users perspective, the following changes are most prominent among other enhancements:
	</para>
	<itemizedlist>
		<listitem>
		<para>
		The data from the module may be send encrypted, with a key stored in secrets.tdb (or secrets.ntdb). The
		Receiver then has to use the same key. The module does AES block encryption over the
		data to send.
		</para>
		</listitem>
		<listitem>
		<para>
		The module now can identify itself against the receiver with a sub-release number, where
		the receiver may run with a different sub-release number than the module. However, as
		long as both run on the V2.x protocol, the receiver will not crash, even if the module
		uses features only implemented in the newer subrelease. Ultimately, if the module uses
		a new feature from a newer subrelease, and the receiver runs an older protocol, it is just
		ignoring the functionality. Of course it is best to have both the receiver and the module
		running the same subrelease of the protocol.
		</para>
		</listitem>
		<listitem>
		<para>
		The parsing problems of protocol V1 can no longer happen, because V2 is marshalling the
		data packages in a proper way.
		</para>
		</listitem>
		<listitem>
		<para>
		The module now potentially has the ability to create data on every VFS function. As of
		protocol V2.0, there is support for 8 VFS functions, namely write,read,pread,pwrite,
		rename,chdir,mkdir and rmdir. Supporting more VFS functions is one of the targets for the
		upcoming sub-releases.
		</para>
		</listitem>
	</itemizedlist>
	<para>
		To enable protocol V2, the protocol_version vfs option has to be used (see OPTIONS).
	</para>
		
</refsect1>		

<refsect1>
	<title>OPTIONS with protocol V1 and V2.x</title>

	<variablelist>

		<varlistentry>
                <term>smb_traffic_analyzer:mode = STRING</term>
                <listitem>
                <para>If STRING matches to "unix_domain_socket", the module will
		use a unix domain socket located at /var/tmp/stadsocket, if
		STRING contains an different string or is not defined, the module will
		use an internet domain socket for data transfer.</para>

                </listitem>
                </varlistentry>


		<varlistentry>
		<term>smb_traffic_analyzer:host = STRING</term>
		<listitem>
		<para>The module will send the data to the system named with
		the hostname STRING.</para>

		</listitem>
		</varlistentry>

		<varlistentry>
		<term>smb_traffic_analyzer:port = STRING</term>
		<listitem>
		<para>The module will send the data using the TCP port given
		in STRING.
		</para>
		</listitem>
		</varlistentry>
		<varlistentry>
		<term>smb_traffic_analyzer:anonymize_prefix = STRING</term>
		<listitem>
		<para>The module will replace the user names with a prefix
		given by STRING and a simple hash number. In version 2.x
		of the protocol, the users SID will also be anonymized.
		</para>

		</listitem>
		</varlistentry>

		<varlistentry>
		<term>smb_traffic_analyzer:total_anonymization = STRING</term>
		<listitem>
		<para>If STRING matches to 'yes', the module will replace
		any user name with the string given by the option 
		smb_traffic_analyzer:anonymize_prefix, without generating
		an additional hash number. This means that any transfer data
		will be mapped to a single user, leading to a total 
		anonymization of user related data. In version 2.x of the
		protocol, the users SID will also be anonymized.</para>
		</listitem>
		</varlistentry>

		<varlistentry>
		<term>smb_traffic_analyzer:protocol_version = STRING</term>
		<listitem>
		<para>If STRING matches to V1, the module will use version 1 of the
		protocol. If STRING is not given, the module will use version 2 of the
		protocol, which is the default.
		</para>
		</listitem>
		</varlistentry>

	</variablelist>
</refsect1>

<refsect1>
	<title>EXAMPLES</title>
	<para>Running protocol V2 on share "example_share", using an internet socket.</para>
	<programlisting>
	<smbconfsection name="[example_share]"/>
	<smbconfoption name="path">/data/example</smbconfoption>
	<smbconfoption name="vfs_objects">smb_traffic_analyzer</smbconfoption>
	<smbconfoption name="smb_traffic_analyzer:host">examplehost</smbconfoption>
	<smbconfoption name="smb_traffic_analyzer:port">3491</smbconfoption>
	</programlisting>

	<para>The module running on share "example_share", using a unix domain socket</para>
	<programlisting>
	<smbconfsection name="[example_share]"/>
	<smbconfoption name="path">/data/example</smbconfoption>
	<smbconfoption name="vfs objects">smb_traffic_analyzer</smbconfoption>
	<smbconfoption name="smb_traffic_analyzer:mode">unix_domain_socket</smbconfoption>
	</programlisting>

	<para>The module running on share "example_share", using an internet socket,
	connecting to host "examplehost" on port 3491.</para>
	<programlisting>
	<smbconfsection name="[example_share]"/>
	<smbconfoption name="path">/data/example</smbconfoption>
	<smbconfoption name="vfs objects">smb_traffic_analyzer</smbconfoption>
	<smbconfoption name="smb_traffic_analyzer:host">examplehost</smbconfoption>
	<smbconfoption name="smb_traffic_analyzer:port">3491</smbconfoption>
	</programlisting>

	<para>The module running on share "example_share", using an internet socket,
	connecting to host "examplehost" on port 3491, anonymizing user names with
	the prefix "User".</para>
	<programlisting>
	<smbconfsection name="[example_share]"/>
	<smbconfoption name="path">/data/example</smbconfoption>
	<smbconfoption name="vfs objects">smb_traffic_analyzer</smbconfoption>
	<smbconfoption name="smb_traffic_analyzer:host">examplehost</smbconfoption>
	<smbconfoption name="smb_traffic_analyzer:port">3491</smbconfoption>
	<smbconfoption name="smb_traffic_analyzer:anonymize_prefix">User</smbconfoption>
	</programlisting>
</refsect1>

<refsect1>
	<title>VERSION</title>
	<para>This man page is correct for version 3.3 of the Samba suite.
	</para>
</refsect1>

<refsect1>
	<title>AUTHOR</title>

	<para>The original Samba software and related utilities
	were created by Andrew Tridgell. Samba is now developed
	by the Samba Team as an Open Source project similar
	to the way the Linux kernel is developed.</para>

	<para>The original version of the VFS module and the
	helper tools were created by Holger Hetterich.</para>
</refsect1>
</refentry>