RE: [squid-users] logs, databases and statistics

From: John Cuson <jcuson@dont-contact.us>
Date: Tue, 21 Aug 2001 12:38:27 -0500

mehdi-

you might want to take a look at the summary i put on the web site ... one of the back-end databases i use is postgresql. in perl, all you have to do to interface with another dbi compliant database is to use the dbd driver for that database. (actually, there's a little bit more to it than that, but i talk about that a little bit in the background document.) the second script would require nothing more than changing the name of the database driver for the source table.

take a look and drop me a line if you have questions.

john cuson
(jcuson@labs.isdh.state.in.us, jcuson@alumni.indiana.edu)

The world is full of obvious things which nobody by any chance ever observes.
                              
                                     Sir Arthur Conan Doyle

>>> Mehdi Melhaa <Mehdi.Melhaa@fr.clientlogic.com> 08/21/01 11:34AM >>>
Hi,

Do you know if the same script exist for postgreSQL database.

Thank you

Regards,

Mehdi

-----Message d'origine-----
De : John Cuson [mailto:jcuson@isdh.state.in.us]
Envoyé : mar. 21 août 2001 17:21
À : squid-users@squid-cache.org
Objet : [squid-users] logs, databases and statistics

hi-

after having installed a squid proxy at this location some months ago i got
to looking around and realized that, though there were scripts that queried
the logs and produced various summaries, there was nothing that i could see
that allowed for more detailed analysis of the access logs. it made sense
to me that the most effective way to do that would be to have a data set
that spanned a reasonable length of time and was capable of supporting a
level of analysis both interactive and iterative. so as a project i started
into looking at it. after developing a little script to automate parsing of
the access log file and inserting the results into a table, i realized that
the nature of the record in the access log is not overly conducive to
statistical analysis, so i started looking at some proxy variables flagging
certain types of dates, and generally creating a transformed record that
would be a better starting point. i've posted the resultant material here:
http://www.geocities.com/ralphie_the_geek/squid.htm .. the page consists of
the two perl scripts that i use to deal with the records, the structure of
the tables, and a two-page summary of the overall system. this is of course
just a starting point, but anyone who wishes to use it is welcome. sorry
for the little geocities ads on the page. what can i say, it's free
<grin>.

john cuson
(jcuson@labs.isdh.state.in.us, jcuson@alumni.indiana.edu)

The world is full of obvious things which nobody by any chance ever
observes.
                              
                                     Sir Arthur Conan Doyle
Received on Tue Aug 21 2001 - 11:38:59 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:01:52 MST