Copyright (c) 1994, 1996 Regents of the University of California. See the file LICENSE for Licensing and Redistribution Information. See the file README for more information. ========================================================================== The latest version of wwwstat can be obtained from http://www.ics.uci.edu/pub/websoft/wwwstat/ ftp://www.ics.uci.edu/pub/websoft/wwwstat/ ========================================================================== Frequently Asked Questions 1. MAILING LIST Due to the popularity of this program, all questions, suggestions, bug reports, fixes, or enhancements should be sent to the wwwstat-users@ics.uci.edu mailing list and not to the author. You can join the list by sending e-mail with "subscribe" in the subject of the message to the request address wwwstat-users-request@ics.uci.edu The list is archived at the distribution site. Send "help" or "info" in the subject of a message to for more information on (un)subscribing. 2. Why is all that legalese in LICENSE necessary? Isn't wwwstat free? The LICENSE exists because others have abused the privilege of using free software. Because this software was developed by an employee of the University of California, we must protect ourselves from lawsuits by those who would abuse our legal system for personal gain, regardless of any actual damages. To our knowledge, no damage has ever been caused by this program. wwwstat is distributed free of charge and will remain so as long as it is legally possible. If you are not distributing the program to others, there is no need for you to include mention of the University of California in its output. However, I would prefer that you leave in the reference to wwwstat's distribution site (at the bottom of the output) so that others can know where to get the original program. wwwstat is in use around the world. If you have translated the output to another language (i.e. German, French, Maori, etc.), I encourage you to share those translations with others by sending the translated configuration file to the mailing list. 3. Will you be developing a version for other httpd's ... Obviously, versions of this program for every HTTP server would be nice. However, I don't have the time to install all those servers just to see how to do it. Feel free to do so yourself. The wwwstat program should work on any system with Perl 4.036 or 5.003 and on any logfile in the Common Logfile Format. It does not need to run on the same system as the httpd server, and will only need network access if the -dns option is being used. 4. I am still using NCSA httpd_1.1 (or earlier), which doesn't support the Common Logfile Format. What can I do? Well, first you should seek professional help. :) Anyone running a version of NCSA httpd prior to 1.4, or CERN httpd prior to 3.0, is running a program with KNOWN SECURITY HOLES. You should upgrade *immediately*, preferably to the Apache server . Version 1.01 of wwwstat included the oldwwwstat script for pre-CLF logfiles. That script is no longer supported, but may be found at the distribution site above. The oldlog2new program is provided for converting old NCSA logfiles (pre-1.2) to the common logfile format. 5. I am using the Apache server with a configurable logfile format. How do I get wwwstat to read the custom logfile? If you know perl, you can always try editing the wwwstat script so that it parses the different format. Seriously, this should be easy to do, provided that the format can be unambiguously parsed. The hard part is doing anything useful with the extra information provided by your custom format. The other option is to write your own conversion script that converts from the custom log format to the CLF. I am sure that I've seen this kind of program before, but I can't find it at the Apache site right now. 6. How do I setup a crontab script to run wwwstat nightly? Well, that depends on how your system's crontab works, but on mine (a Sun 4 running SunOS 4.1.2) I can edit the crontab with the command % crontab -e I have the following entry for my nightly script: 15 0 * * * /dc/ud/www/etc/update-stats and the following is my update stats script (thanks to Hal Varian) ---------------------------------------- #!/bin/sh /dc/ud/www/bin/wwwstat -D 'today' > /tmp/wwwstats.html mv -f /tmp/wwwstats.html /dc/ud/www/documentroot/Admin/wwwstats.html exit ---------------------------------------- 7. What is the general procedure for monthly resetting of the access_log? Again, that depends a great deal on how your site is set up and how frequent the accesses are to your server. My site gets about 150000 requests a month, so I have a perl script (monthly.pl) which I run a couple days after the beginning of each month. monthly.pl reads the logfile (assumed to contain more than one month's worth of WWW common logfile entries) and moves the prior month's entries into a separate file. The new file is created on TMPDIR (to avoid filling up the disk), compressed using gzip, and then moved to the archive directory. The program also restarts the httpd server. The monthly.pl script is provided AS AN EXAMPLE. You must configure it (by editing the script) before it can be safely used. Many sites will have too large a server load for only monthly resetting of the logs. In that case, modify the scripts so that they do things on a weekly or daily basis. Sorry, I do not have examples of that. 8. My server load is HUGE and wwwstat runs out of memory, what can I do? You can use the -trunc N option and/or the -nofiles option to reduce the memory consumed for the archive section. You can also reduce memory consumed by configuring special groupings using the %DomainMap and @ArchiveMap tables. Finally, you can exclude entire sections using the -no
options (see the man page). If all that fails, then just run wwwstat on less input. 9. What is perl? Where can I get it? Does perl work for my platform? See . *. MORE ANSWERS The above is just what I knew about prior to the current distribution. More questions and (hopefully) answers can be found at the FAQ page linked from the distribution site.