|
 |
Solutions hébergement |
 |
Support Webalizer. |
| |
|
Source : http://www.mrunix.net/webalizer/
|
|
1. The Webalizer doesn't show me Referrers or User Agents?
In order for the Webalizer to produce statistics for user agents (browsers)
and referrers, that information needs to be in the log files produced by the
web server. Most servers by default only produce CLF logs, which do not
include the extra information. The way you have your server include this
information depends on what server you are running. For apache, you
need to edit the httpd.conf file (in the servers /conf directory) and...
For apache 1.2, add the line:
LogFormat "%h %l %u %t \"%r\" %>s %b \"%{Referer}i\" \"%{User-agent}i\""
For apache 1.3, use the line:
CustomLog /var/lib/httpd/logs/access_log combined
Other servers are similar. Refer to your servers documentation for
additional information on how to enable referrers and user agents.
2. What systems will The Webalizer run on?
The Webalizer was designed on and for an Intel system running Linux,
however was written to be as close to ANSI/POSIX specs as possible in
order to be easily ported to other platforms. I currently only have
access to Linux systems running on Intel and PowerPC hardware, so can
verify that it runs on those. In addition, I have received
lots of mail from users indicating that The Webalizer will run on just
about any *NIX machine out there from AIX to XENIX. (Other platforms
also supported, such as OS/2 and MacOSX.. check the download page).
3. I get 'libgd not found' errors when
compiling?
The Webalizer uses the gd graphics library written by Tom Boutell for
producing it's inline graphics. If you don't have this library or have
it installed correctly, you will get this error. The Webalizer expects this
library to be in the standard library path (ie: /usr/lib), so if you have it
someplace else, you need to add an '-L[path]' flag to CFLAGS in the Makefile.
If you don't have the library, get it at
4. I get 'No File or Directory' errors when
compiling?
The Makefile supplied with The Webalizer expects to find the header files
for the gd graphics library in /usr/local/include/gd. If they are located
somewhere else, you can either create a symbolic link to them, or edit
the Makefile and specify the correct location.
5. What is the difference between 'HITS' and 'FILES'?
There is an on-line quick help that
explains most of the major terms and headings used in the reports. Basically,
HITS is the total number of HTTP requests that the server received
during the reporting period. Any request made to the server is considered
a hit. FILES is the number of hits that actually resulted in
something being sent back to the user, such as an HTML page or image.
'Total Files' and '200 - OK' totals should be the same. If you add up
the totals in the 'Hits by Response Code' section, it should be the same
as the 'Total Hits' figure.
6. My logs are HUGE! Can I run The Webalizer on partial logs?
Yes!. With the release of version 1.2x, The Webalizer
now supports incremental processing. This allows you to rotate your
logs as often as needed without the loss of statistical detail between
runs. Use the "Incremental" keyword in your configuration file, or the
"-p" command line switch to enable incremental processing.
7. My Server doesn't do name lookups. Will The Webalizer?
Yes. Version 2.00 and higher supports reverse DNS support. If you don't enable hostname lookups on
your web server, you will get '100% Unresolved/Unknown' country
totals. This is because your log files only have IP addresses and not
names. While it is recommended that you let your web server handle the
DNS lookups, DNS support can be used for those sites where DNS resolution
is not an option.
8. I used the [Hide*] option, but it still shows up in the totals?
Using the Hide* options only prevent that object from
being displayed in the 'Top' table generated by The Webalizer. It is still
counted in the totals. Version 0.99 of The Webalizer now has
Ignore* options, which allow you to completely ignore
certain objects for statistical purposes.
9. I used the [Group*] option, but it still shows up?
Grouped items, by default, are not hidden. This allows you to display
a group total as well as all the items that make up the grouping. If
you don't want to see the individual items that match the group, then
follow the Group* keyword with an identical Hide* one.
10. Changing the configuration file has no effect?
Which configuration file are you changing? The Webalizer looks in the
current directory for a file named webalizer.conf, which it will
process before any other configuration files. If one is not
present in the current directory, it will look for the file
/etc/webalizer.conf, and process it before any other configuration
files. Some configuration options allow you to toggle settings on or
off, while some cannot be reversed. If you, for example specify the
configuration option "HideURL *.gif" in the system wide default
file /etc/webalizer.conf, you cannot 'un-hide' that object using
a local configuration file. In general, single sites should have a
single configuration file, such as /etc/webalizer.conf. Larger
sites that have multiple host/virtual hosts probably should use different
configuation files for each host and not have a default
webalizer.conf file.
11. My configuration file is being read twice!
Do Not use '-c webalizer.conf' on the command line! This
file is always read if found, regardless of any other configuration
files that may be specified. If you do specify it on the command line, it
will be read twice...
12. I get Error adding xxx node, skipping ... errors! Why?
You ran out of memory. The error occurs when a malloc call is made to
allocate free memory, and fails. You can increase your swap space, but the
only real solution is to add more physical memory.
13. I get Warning: Truncating oversized xxx or
String exceeds storage size warnings! Why?
Internally, The Webalizer has a fixed maximum size for various
parts of the log record. If a particular field is longer than will fit, you
will see these warnings. The most common is that for the request field
on sites that have a lot of CGI interaction. They can be safely ignored. If
you don't want to see warnings or errors, you can use the ReallyQuiet option
(-Q command line switch) to supress them.
14. Why don't the daily visit totals add up to the monthly total?
You cannot add up the daily visit totals and compare them to the monthly
total, they are different reporting periods. For example, if someone visits
your site at 11:45pm and stays until 12:15am, the monthly total would show one
visit, while the daily totals will show two (one for each day).
15. Why do my reports show more Sites than
Visits?
Visits are only triggered when a valid request is found for a
page, as defined by your PageType setting (or a URL that
ends with a slash, which is also considered a page type). Sites however,
are counted regardless of the request type. It is very common to have
more sites than visits, particularly if you host non-pagetype URLs on
your site that are linked to from the outside. If you are not hosting
URLs that are linked to from outside sites, then make sure your PageType
setting is correct. The default is .htm, .html and .cgi extensions, unless
you specify otherwise.
16. How can I process multiple logs from a server farm?
Q:
I have multiple load-balanced servers (or I'm using DNS round-robin to
accoplish the same thing) and I want to generate one webalizer report for
the whole farm, but each server generates its own log file. When I
run webalizer on each of the logfiles in turn, it ignores a lot of the
records because it thinks they're out of order!
A:
You need to merge all of the logfiles together so that webalizer sees
the records in chronological order. One good way to do that on the fly is
with mergelog
(
http://mergelog.sourceforge.net/)
, a quick common logfile sorter. An example:
mergelog .log .log .log | webalizer
Another method is to simply combine your logs and then sort them into
chronological order. Here is a simple shell script that uses the GNU
sort utility to sort an already combined log file:
#!/bin/sh
if [ ! -f $1 ]; then
echo "Usage: $0 "
exit
fi
echo "Sorting $1"
sort -t ' ' -k 4.9,4.12n -k 4.5,4.7M -k 4.2,4.3n -k 4.14,4.15n -k 4.17,4.18n -k 4.20,4.21n $1 > $1.sorted
17. How can I easily process multiple virtual hosts?
There are many ways to process multiple virtual hosts on the same
machine. The easiest way I have found, provided that each host
generates it's own log file, is as follows:
- Create a central directory for your configuration files.
(I use /etc/webalizer)
- Make a configuration file for each virtual host and place them
in the central directory. Each configuration file should have
at least the HostName (domain), OutputDir and
LogFile configuration settings specified. You probably
will want to specify other settings specific to the domain,
such as HideReferrer, HideSite and maybe some
others as well. Name the file the same as the domain name,
and end it with a .conf extension, so you can easily tell
what vhost the configuration is for.
- To process all your virtual sites with a single command, a
simple shell command can now be used:
for i in /etc/webalizer/*.conf; do webalizer -c $i; done
After you have it set up, to add a new host, all you need to do
is create a new configuration file and put it in the directory.
It will be automatically picked up the next time you run the
command.
18. I am having problems compiling w/DNS support!
Lots of people have problems compiling DNS support into the Webalizer,
which is why you have to specifically enable it. If you don't really
need DNS support, don't try to compile it in. The vast majority of
sites can get by simply turning on hostname lookups on their web
server, which will do the DNS resolution for you automatically.
If you really need built in DNS support, it's really quite simple,
however different distributions place the required headers and
libraries in different places, so the configure script fails to
find them. In a nutshell, you need the
Berkeley DB
libary, and it needs to be configured with V1.85 API support.
Most Linux distributions already have this library present.
If you get errors about not finding the proper header file (db_185.h),
locate it on your system and create a symbolic link to it in the
/usr/include directory. You will also need to specify the correct
library to use for the header. This may mean that you need to first
run the configure script, then if the compile fails due to unresolved
references, edit the Makefile and change the -ldb (or -ldb1)
reference to the correct library, such as -ldb-3.2 for RedHat, or
just -ldb for Slackware 8. Each distribution either names the
library something different, or puts it in a different location, so you
will have to play around with it on your distro to get it to work.
If the library is in some non-standard location, you may also need to
specify it's path using the --with-dblib=... switch when you
run the configure script. Bottom line is, if you don't really need
it (and most people don't), just compile without DNS support and let
your web server do the name lookups for you.
Submitted by Victor Brilon:
To compile under RH 7.x or 8.x using DNS resolver, you must use this
command line:
./configure --enable-dns --with-db=/usr/include/db1
You must have the db1 and the db1-devel RPMs installed in order to do this.
19. Where can I get additional help?
First, read the documentation! 99.9% of all questions I receive are
answered in the available documentation. Start by carefully reading
the README
file . If you still can't
find an answer, try doing a quick search on the web using your favorite
search engine, or in the newsgroups. Installation problems can usually
be solved by reading the
INSTALL
document, or the many third party documents found on the web, such as
Danilo C. Dy's
documentation packages (look under the SysAdmin Guides section).
He has an
Alternate Site here.
There is also a discussion forum at
Yahoo Groups.
If all else fails, you can send me an e-mail, however due to the amount
of mail I receive each day, you may or may not get a timely response.
|
|
|
Nos serveurs
sont désormais des serveurs
Pentium 3 Ghz, 1 Go Ram
|
La
formule de base est à 10
Euros TTC / mois
Si vous avez des besoins plus spécifiques (composants,
espace disque...), nous sommes la pour répondre à
vos questions. |
Rappel : les
frais d'installation sont gratuits |
|
|
|