 |
Chapter 26 |  |
|
Chapter 26 | |
WebSite has three main logs: the access log, the error log, and the server log. The access log records each request to the server and the server's response in one of three formats: Common (older NCSA/CERN), Combined NCSA/CERN, and Windows Log format. The error log records access errors, such as failed user authentication. The server log records each time the server is restarted or its configuration is updated. You can also set a variety of tracing options to have the server log collect more or less detailed information. The tracing information is used primarily for troubleshooting and debugging.
WebSite log configuration occurs on the Logging page of Server Admin. shows the default Logging page.
![[Graphic: Figure 26-6]](figs/wm_2406.gif)
The Log File Paths section allows you to choose the path and filenames of the log files. The default files are in the directory /WebSite/logs. You can place a full pathname to put the logs in another location. The server will shut down if it cannot maintain log files. Therefore, if you locate the logging directories on a remote computer, make sure that it will always be available when the server is running.
The Access Log Format section of the page allows you to select the format of the access log file. The three formats are described as follows:
This format collects server access data in a format that can be easily imported into most Microsoft Windows Office productivity packages. The entries are tab delimited and require no additional parsing using Visual Basic or Perl. This format creates larger files than the other logging formats.
This format collects server access data in a more standard Web log format with fields delimited by quotation marks. Two extra fields are included to identify the URL from which the browser made the current request (Referer), and to identify the browser type (User_Agent).
This is the default format for WebSite and is the standard Web log format used by most Web servers.
Instead of the IP address, you can have WebSite find and put the requesting client's hostname in the log files. To do this, select the Enable DNS Reverse Lookup box in the Client Hostname Lookup section of the Logging page. For every logged access, the server will contact its DNS server with the browser's IP address and ask for its corresponding DNS hostname. It is not suggested that you use this feature without a compelling reason as it will slow transaction time noticeably.
WebSite has 10 different tracing options that provide detailed data about the server's activity in the server log. You can select one or more of these options by clicking on the appropriate checkboxes in the Tracing Options sections of the Logging page. The first six options listed are useful to the server administrator, while the last five are used mainly by technical support. The last five will not be described here. You can remove any tracing option by clicking on a checked box, or press Clear All Tracing to remove them all. The first six tracing options are as follows:
This tracing option records the incoming header data for each request from a browser and the action the server takes responding to the request. See Chapter 17, HTTP Overview through Chapter 20, Media Types and Subtypes for more information on HTTP.
This tracing option records all of the outgoing data to the browser in the server's response (header and file data). This option generates huge server logs.
This tracing option records the information about a client requesting a location on a clickable image map, and shows the server's response.
This option records the server's activity when a browser requests a URL containing a CGI or WSAPI execution. It is useful for debugging CGI programs. When this tracing option is enabled, the server notifies the various types of CGI or API programs to enable their own debugging features, such as displaying standard output and saving temporary files. See Chapter 9, CGI Overview for more information on CGI.
This tracing option records the server's actions in checking access control restrictions and then denying or allowing access to a specific URL path by requestor. Access Control tracing shows both the class restrictions (IP address or hostname) and user authentication requirements. User and group names are not included with this tracing option.
This tracing option records all user authentication attempts, whether successful or unsuccessful, and why they weren't. Passwords are shown only in encrypted form. The URL path for the authentication attempt is not included with this tracing. Enabling both Access Control and Authentication tracing provides a complete picture of how your server is handling access control restrictions.
