NAME

sarg - Squid Analysis Report Generator

SYNOPSIS

sarg [options] [logfile...]

DESCRIPTION

sarg is a log file parser and analyzer for the Squid Web Proxy Cache[1]. It allows you to view "where" your users are going to on the Internet.

sarg generates reports in HTML with fields such as: users, IP Addresses, bytes, sites, and times. These HTML files can appear in your web server's directory for browsing by users or administrators. You may also have sarg email the reports to the Squid Cache administrator.

sarg can read squid or Microsoft ISA access logs. Optionally, it can complement the reports with the log of a Squid filter/redirector such as squidGuard[2].

OPTIONS

A summary of options is included below.

-h --help

Show summary of options.

-a hostname|ip address

Limits report to records containing the specified hostname/ip address

-b filename

Enables UserAgent log and writes it to filename.

-c filename

Read filename for a list of the web hosts to exclude from the report. See the section called “HOST EXCLUSION FILE”.

--convert

Convert a squid log file date/time field to a human-readable format. All the log files are read and output as one text on the standard output.

--css

Output, on the standard output, the internal css sarg inlines in the reports. You can redirect the output to a file of your choice and edit it. Then you can override the internal css with external_css_file in sarg.conf.

Using an external css can reduce the size of the report file. If you are short on disk space, you may consider exporting the css as explained above.

-d date

Use date to restrict the report to some date range during log file processing. Format for date is dd/mm/yyyy-dd/mm/yyyy or a single date dd/mm/yyyy. Date ranges can also be specified as day-n, week-n, or month-n where n is the number of days, weeks or months to jump backward. Note that there is no spaces around the hyphen.

-e email

Sends report to email (stdout for console).

-f filename

Reads configuration from filename.

-g e|u

Sets date format in generated reports.

e = Europe -> dd/mm/yy

u = USA -> mm/dd/yy

-i

Generates reports by user and ip address.

Note
This requires the report_type option in config file to contain "users_sites".

--keeplogs

Don't delete any old report. It is equivalent to setting --lastlog 0 but is provided for convenience.

-l filename

Uses filename as the input log. This option can be repeated up to 255 times to read multiple files. If the files end with the extension .gz, .bz2 or .Z they are decompressed. If the file name is just -, the log file is read from standard input. In that case, it cannot be compressed.

This option is kept for compatibility with older versions of sarg but, starting with sarg 2.3, the log files may be named on the command line without the -l option. It allows the use of wildcards on the command line. Make sure you don't exceed the limit of 255 files.

--lastlog n

Limit the number of logs kept in the output directory to n. Any supernumerary report is deleted starting with the oldest report. The value of n must be positive or zero. A value of zero means no report should be deleted.

-L filename

Reads a proxy redirector log file such as one created by squidGuard or Rejik. If you use this option, you may want to configure redirector_log_format in sarg.conf to match the output format of your web content filtering program. This option can be repeated up to 64 times to read multiple files.

-n

Enables ip address resolution.

-o dir

Writes report in dir.

-p

Generates reports using ip address instead of userid.

-P prefix --splitprefix prefix

This option must be used with --split. If it is provided, the input log is split among several files each containing one day. The name of the output files is made of the prefix and the date formated as -YYYY-MM-DD.

The output files are written in the output directory specified with -o or in the current directory.

-r

Output the realtime report on the standard output and exit.

-s string

Limits report to the site specified by string [eg. www.debian.org]

--split

Split the squid log file and output it as text on the standard output omitting the dates outside of the range specified by the -d parameter. If it is combined with --convert the dates are also converted to a human-readable format.

Combined with -P, the log is written in several files each containing one day of the original log.

--statistics

Writes some statistics about the execution time. The statistics include the total execution time; the number of records read in the input log files and the time it took to read them; the number of records and users processed and the time it took to process them.

-t string

Limits the records included in the report based on time-of-day. Format for string is HH:MM or HH:MM-HH:MM. The former reports only the requested time. The latter reports any entry falling within the requested range. This limit complement the limit imposed by option -d.

-u user

Limits reports to user activities.

-v

Write sarg version and exit.

-w dir

Store temporary files in dir. In fact, sarg stores its temporary files in the sarg subdirectory of dir. Be sure to set the HTML output directory to a place outside of the temporary directory or sarg may fail or delete the report when it completes its task.

-x

Writes debug messages to stdout

-z

Writes process messages to stdout.

HOST EXCLUSION FILE

Sarg can be told to exclude visited hosts from the report by providing it with a file containing one host to exclude per line. The "host" may be one of the following:

•a full host name,

•a host name starting with a wildcard (*) to match any prefix,

•a single ip address,

•a subnet noted a.b.c.d/e.

Example 1. Example of a hosts exclusion file

*.google.com

10.0.0.0/8

Sarg cannot exclude IPv6 addresses at the moment.

AUTHORS

This manual page was written by Luigi Gangitano<gangitano@lugroma3.org>, for the Debian GNU/Linux system (but may be used by others). Revised by Billy Newsom.

Currently maintained by Frédéric Marchal<fmarchal@users.sourceforge.net>.

AUTHORS

Frédéric Marchal <fmarchal@users.sourceforge.net>

Docbook version of the manual page

Billy Newsom

Revision of the manual page

Luigi Gangitano <gangitano@lugroma3.org>

Author of the first manual page

COPYRIGHT

NOTES

1.: Squid Web Proxy Cache

http://www.squid-cache.org/

2.: squidGuard

http://www.squidguard.org/