TUX Log files format and scripts for analysis of log files

[lisbon_pt]

Hi

I have the following scenario (LINUX):

tux and apache working in the same DOCROOT

I use:
      tux2w3c   native_log_file_tux  >  my_tux_logs_in_txt

The lines in my_tux_logs_in_txt  have the following format
(which is W3C compatible):

194.65.122.241 - - Mon Feb 27 03:50:42 2006 "GET
site.pt/somedir/images/IMAGE.gif HTTP/1.0" - 0 404

I have tried scripts (analog, calamaris, awstats) to analyse these logs
(my_tux_logs_in_txt) defining a LOGFORMAT

For example for analog, I've tried:
LOGFORMAT (%S %j %u %j %M %d %h:%n:%j %Y "%r" %j %b %c)

for the line logs like the ones:

194.65.122.241 - - Mon Feb 27 03:50:42 2006 "GET
site.pt/somedir/images/IMAGE.gif HTTP/1.0" - 0 404


The outpur of analog is:

This report contains overall statistics.
Successful requests: 0
Failed requests: 21,894
Corrupt logfile lines: 225
--------------------------------------------------------------------------------
Status Code Report
(Go To: Top: General Summary: Status Code Report)
This report lists the HTTP status codes of all requests.
Listing status codes, sorted numerically.

 reqs: status code
-----: -----------
21894: 404 Document not found


I take for granted:

1. Tux is not working has cache (Is LOGFORMAT well defined?)
2. Why the corrupted log lines in log file?

Thanks very much for your (precious) time reading and maybe providing
some feedback.

Best regards,
lisbon_pt

-------------- ANALOG for LOGFORMAT --------------
%S  host (the client hostname, or address of the computer making the request)
%r   file requested
%u  user (tip: a cookie or session id can usefully be defined as %u too)
%d  day of the month
%M  month, three letter English abbreviation
%Y   year, four digits
%h   hour of the day
%n   minute of the hour
%b   number of bytes transferred
%c   HTTP status code
%j   junk: ignore this field (field can be empty too)