Apache at your Web Service

You are here: irt.org | Articles | Linux & Apache | Apache at your Web Service

Published on: Sunday 1st August 1999 By: Pankaj Kamthan

• Introduction
• Apache User-Base
• Features of Apache
• Advantages of Apache
• Obtaining Apache
  • Binary Vs. Source
• Apache on Specific Operating Systems
  • UNIX
    • Compiling Apache on UNIX
    • Installing Apache on UNIX
    • Running Apache on UNIX
    • Linux
  • Windows
    • Compiling Apache on Windows
    • Installing Apache on Windows
    • Running Apache on Windows
  • Macintosh
• Log Files
• Troubleshooting
  • Compilation Problems
  • Installation Problems
  • Running Problems
• Support
• Limitations of Apache
• Acknowledgements
• Conclusion
• References
• Appendix : Apache Command Line Options

Introduction

Apache began its life in 1995 as an enhanced version of the NCSA Web server through a series of "patches" and the result was an APATCHY server. Over the years Apache diverged from the original as many features were added to it. The main legacy in it are the configuration files, which have been kept for backward compatibility with the NCSA Web Server.

Since its inception, Apache has reached many landmarks. In 1997, Apache ports for Windows 95 and NT platforms were announced. In 1998, IBM announced support for the Apache project and a version of Apache is planned in its Internet commerce servers.

A detailed account of how Apache came about is available at www.apache.org. The current distributed version is 1.3.x, and forms the basis of this article.

Apache User-Base

The benchmark comparisons provided by Netcraft surveys show that the use of Apache has grown steadily over the last five years, and it is the most popular among the Web servers in use today. As of July 1998, Apache is run on over 1.2 million Internet servers.

Figure 1 illustrates the statistics on server share for the period August 1995 - May 1999. You may notice the (horizontal) symmetry indicating the rise of Apache and corresponding decline of NCSA over this period of time.

Figure 1. Netcraft Survey on Server Share.

Features of Apache

A comparison of features of a large number of Web servers currently in use is available at WebServer Compare. A quick overview of features of Apache is available at webcompare.internet.com. A review of Apache by Server Watch is available at serverwatch.internet.com.

Advantages of Apache

Apache offers various advantages to users, developers and Web administrators:

• Features. Apache has various useful features, including implementation of the latest protocols.
• Customizable. Apache's modular architecture allows you to build a server that is "made to measure."
• Administration. Apache configuration files are in ASCII, have a simple format, and can be edited using any text editor. They are transferable, so one can effectively clone a server. One can control the server from command line, which makes remote administration very convenient.
• Extensible. Apache server and API source code are open to public. If there is any feature that you want but does not exist in Apache, you can write your own server module to implement it.
• Efficient. A lot of effort has been put into optimizing the Apache's C code for performance. As a result, it runs faster and consumes less systems resources than many other servers.
• Portability. Apache runs on a wide variety of operating systems, including all variants of UNIX, Windows 9x/NT, MacOS (on PowerPC), and various others.
• Stability/Reliability. Apache's source code is open to public. When any bugs are found, they are often quickly communicated, and rapidly fixed. Updates are made and announced thereafter. This has resulted in Apache becoming more and more stable, and hence reliable, server over the time.
• Support. Apache is supported by the Apache Group, a large number of dedicated users (many of whom can often be found in the Usenet newsgroup comp.infosystems.www.servers.unix or comp.infosystems.www.servers.ms-windows), and by companies who market commercial versions of Apache.

Obtaining Apache

Apache is available as source as well as in binary formats for various operating systems. These sites list the current release, as well as, recent beta releases. Current distributed version is 1.3.6. Apache is mirrored world-wide. The best way to download Apache is to find the mirror closest to you.

Binary Vs. Source

There is an inherent trade-off between installing a binary and (compiling and) installing a source. A binary, although easier to install, may not necessarily be optimized for your system. A source is relatively more time consuming and difficult to install, as one has to follow a series of steps and make sure that appropriate versions of compilers are available on the system. However, the advantage of that is one can much better customize the server according one's needs.

Apache on Specific Operating Systems

In this section, we will give the details of compiling, installing and running Apache on UNIX and Windows, and briefly discuss the Apache port for Macintoshes.

UNIX

By UNIX, we will mean UNIX and its variants, including Linux. All Apache administration on UNIX is command-line based.

Compiling Apache on UNIX

Change into the src directory of the Apache distribution. Compiling Apache consists of the following three steps:

1. Select modules to compile into Apache in the Configuration file. A detailed account of the significance of Apache modules and their customization is given in Made to Measure : Apache Web Server Customization.
2. Configure Apache for your operating system. Run the Configure script. This generates a Makefile for use in step 3. (If this fails or you have any special requirements, you might need to edit one or more of the following options in the Configuration file: EXTRA_CFLAGS, LIBS, LDFLAGS, INCLUDES.)
3. Type make.

You should now have a binary file called httpd in the src directory. A binary distribution of Apache will supply this file.

Installing Apache on UNIX

Once you have the Apache binary, either as a result of compilation or from download, the next step is to install the program and configure it. Apache is designed to be configured and run from the same set of directories where it is compiled. We now focus our attention to editing the configuration files for the server.

Configuring Apache on UNIX

There are three central configuration files which contain various directives that need to be set up. By default, these files are located in the conf directory and are called srm.conf, access.conf and httpd.conf. The names of these files are taken to be relative to the server root which is set by the ServerRoot directive, the -d command line flag, or (on Windows only) the registry. The files include helpful comments and description of the directives. These files should be edited with care; failure to set up these files correctly could lead to the server not working or being insecure. They could be edited in the following order:

1. httpd.conf. This file contains directives that control the operation of the server daemon. It sets up general attributes about the server: the port number, the user it runs as, and so on.
2. srm.conf. This file contains directives that control the specification of documents that the server can provide to clients. This sets up the root of the document tree, special functions like server-parsed HTML or internal imagemap parsing, etc.
3. access.conf. This file contains directives that control access to documents.

The server also reads a file mime.types containing MIME document types. All these file names can be overridden by appropriate directives.

Running Apache on UNIX

On UNIX, the httpd program is usually run as a (HTTP) daemon which executes continuously, handling requests.

Starting the Server

To start the server, simply run httpd. This will look for httpd.conf in the location compiled into the code (by default /usr/local/apache/conf/httpd.conf). If this file is somewhere else, you can give the real location with the -f option. For example:

/usr/local/apache/httpd -f /usr/local/apache/conf/httpd.conf

If all goes well, this will return to the command prompt almost immediately which indicates that the server is now up and running. If anything goes wrong during the initialization of the server, you will see an error message on the screen.

If the server started properly, you can now use your browser to connect to the server. Once Apache starts running, it listens to port 80 (unless you changed the Port, Listen or BindAddress directives in the configuration files). To connect to the server, start a browser (on the same machine as the server) and enter the URL:

http://localhost/

or if your host isn't connected to the Internet:

http://127.0.0.1/

This should respond with a "test page" as shown in Figure 2, which is a sign of successful installation.

Figure 2. Apache Test Page.

When the server starts, it will create a number of child processes to handle the requests. If you started Apache as the root user, the parent process will continue to run as root while the children will change to the user as given in thehttpd.conf file.

If you want your server to continue running after a system reboot, you should add a call to httpd to your system start-up files (typically rc.local or a file in an rc.N directory). This will start Apache as root. Before doing this, ensure that your server is properly configured for security and access restrictions. The process ID of this process is written to the file httpd.pid in the log files directory logs (unless configured otherwise).

Stopping the Server

A typical command to stop the server is:

kill -TERM `cat /usr/local/apache/logs/httpd.pid`

Restarting the Server

Sending the HUP signal to the parent process causes it to kill off its children like in TERM but the parent doesn't exit. It re-reads its configuration files, and re-opens any log files. Then it spawns a new set of children and continues serving pages.

There is also a script src/support/apachectl which, after a bit of customization for the system, can be used to start, stop, and restart Apache.

Linux

Apache comes bundled with, for example, the RedHat Linux distribution CD-ROM. It can be installed as a service during the installation of the operating system itself. The version on the CD, however, may not always be the latest. The easiest way to upgrade Apache on Linux to the latest version, is to use the RPM. There is an Apache RPM for Linux available and is maintained regularly.

To check whether Apache is running on your system, you can type the URLhttp://localhost/ (or http://127.0.0.1/) in your browser. For details on the Apache httpd, you can type

ps -ef

on the command line.

Table 1 of the Appendix gives the command line options for administering Apache.

Windows

Apache 1.3 is originally designed to run on Windows NT 4.0. An installation test was carried out successfully by the author on Windows 9x and NT 4.0 (Service Pack 4). Note that the binary installer only works on Intel processors. In all cases, TCP/IP networking is required.

Compiling Apache on Windows

Compiling Apache requires (at least) Microsoft Visual C++ 5.0 to be properly installed. We will not discuss the details of the compilation here.

Installing Apache on Windows

When you run the Apache binary installer, you will be asked for:

• The directory to install Apache into.
• The Start menu name.
• The installation type. The "Typical" option installs everything except the source code. The "Minimum" option does not install the manuals or source code. Choose the "Custom" install if you want to install the source code.

This results in the following (if the default settings as used) to be attached in the Start menu:

During the installation, Apache will configure the files in the conf directory for corresponding to the installation directory. After installing Apache, you should further edit these files and adapt it to your needs.

Configuring Apache on Windows

Apache is configured using the files in the conf directory. These are the same as files used to configure the UNIX version, with a few different directives. The main differences are:

• Apache for Windows is multithreaded. So the process-management directives are different:
  • MaxRequestsPerChild. This controls how many requests a process will serve before exiting. Since a process serves all the requests at once, it is recommended that a very high number is used, if this is set. The recommended default is 0 (the process never exits).
  • ThreadsPerChild. This is the maximum number of connections the server can handle simultaneously. This should be set to a high number if it is anticipated that the site will have a lot of traffic. The recommended default is 50.
• The directives that accept filenames as arguments use Windows-style file names. However, since Apache uses UNIX-style names internally, you must use forward slashes. Drive letters can be used; if omitted, the drive with the Apache executable is assumed.
• Apache for Windows contains the ability to load modules at runtime, without recompiling the server. When Apache is compiled, it installs a number of optional modules in the \Apache\modules directory. To activate these, or other modules, the LoadModule directive must be used.

See the Apache documentation for all the available directives.

Running Apache on Windows

On Windows, Apache is usually run as a console application on Windows 9x, or as a service on Windows NT.

Running Apache on Windows as a Console Application

This option is available for Windows 9x and NT. To run Apache from a console window, select the Start Apache as console app option from the Start menu. This will open a console window and start Apache running inside it. The window will remain active until you stop Apache. To stop Apache running, either select the Shutdown Apache console app from the Start menu or enter Control-C in the Apache window. Figure 3 shows Apache running on Windows 9x/NT on console.

Figure 3. Apache running on Windows 9x/NT.

The Start menu icons can provide a simple interface for administering Apache. In some cases, however, it is easier to work from the command line. Table 2 of the Appendix gives the command line options for administering Apache.

Running Apache on Windows as a Service

This option is available on Windows NT only. This is the best choice if you want Apache to automatically start when you boot the machine, and to keep Apache running when you log off.

To start Apache as a service, you first need to install it as a service. Multiple Apache services can be installed, each with a different name and configuration. To install the default Apache service named "Apache", run the Install Apache as Service (NT only) option from the Start menu. Once this is done, you can start the "Apache" service by opening the Services window (in the Control Panel), selecting Apache, then clicking on Start. Apache should now be running in the background. You can later stop Apache by clicking on Stop from the Services window.

Once Apache starts running (either in a console window or as a service), it listens to port 80 (unless you changed the Port, Listen or BindAddress directives in the configuration files). To connect to the server, start a browser (on the same machine as the server) and enter the URL:

http://localhost/

or if your host isn't connected to the Internet:

http://127.0.0.1/

This should respond with a "test page" as shown in Figure 2.

The Start menu icons and the NT Service manager can provide a simple interface for administering Apache. In some cases, however, it is easier to work from the command line. Table 3 of the Appendix gives the command line options for administering Apache.

Macintosh

WebTen from Tenon Intersystems, is a commercial Apache-derivative Web server for Power Macintoshes. WebTen has a browser-based interface and supports standard Macintosh-style plug-ins, CGI, and SSL. Currently, WebTen for MacOS and iTools (previously WebTen v4.0) for MacOS X Server are available.

Log Files

Apache comes equipped with log files (in the logs directory) which contain information regarding the activity of the server.

• Process ID Log. When started, Apache saves the process ID of the parent process in the file httpd.pid. This process ID can be used for restarting and terminating the server.
• Error Log. Apache logs error messages in a log file, error_log, on UNIX and error.log on Windows 9x/NT. This file can be used for monitoring server activities, particularly start-up problems or other failed efforts by the server.
• Transfer Log. Apache typically logs each request to a transfer file, access_log, on UNIX and access.log, on Windows 9x/NT. This file can be used for monitoring user access, and gathering basic information about the user's environment.

All these file names can be overridden by appropriate directives.

Troubleshooting

There are various problems that can occur in getting Apache to work. They are usually related to system or compiler incompatibilities, or improper settings in the configuration files.

Compilation Problems

If the server source doesn't compile on your system, it could be due to any one of the following reasons:

• Operating System Problems. Your operating system could be incompatible and needs to be upgraded.
• C Compiler Problems. Some operating systems include a default C compiler that is either not ANSI C-compliant or suffers from other deficiencies. The recommendation, in such a case, is to install gcc and compile the server with it. After any compiler installation/upgrade (or operating system upgrade), you should make sure that your C header files and include directory tree are consistent with the compiler (and the operating system).

Installation Problems

If a server binary doesn't install on your system, it could be due to any one of the following reasons:

• You downloaded a binary incompatible with your system and tried to install it.
• There are system-specific problems (such as insufficient space or memory) for which you will get an error response, and you can take appropriate steps.

Running Problems

• If the server complains about being unable to "bind" to an address, then either some other process is already using the port you have configured Apache to use, or the server is running as a "normal" user but trying to use a port below 1024 (such as the default port 80).
• If the server is not running at all, then read the error message displayed when you try to run it and also check the server error_log (error.log) file for additional information.

Support

Most problems that get reported to The Apache Group are recorded in the bug database that you can refer to. Another useful resource is the The Apache FAQ.

Depending on your platform, Usenet newsgroups comp.infosystems.www.servers.unix for UNIX and its derivatives or comp.infosystems.www.servers.ms-windows for Windows 9x/NT, can be quite helpful if you have installation/configuration questions. Both are frequently visited by experienced users and you can usually expect a quick response to your query.

Limitations of Apache

Even though Apache is an extremely powerful server with a wide assortment of features, it does have certain, arguably minor, caveats:

• Apache does not provide the GUI-based convenience administration of configuration files.
• The Macintosh version of Apache (unlike its UNIX or Windows counterparts) is not freely available.
• Apache (except the Windows NT version) is not multithreaded.
• The Windows versions do not have all the bells and whistles found in the UNIX versions. For example, it has not yet been optimized for performance. As the Windows version is being actively pursued, it is expected that these differences will vanish in future.

Conclusion

The decision of choosing the "right" Web server is crucial for anybody who intends to have a presence and provide a service on the Web. Apache, with its numerous features, makes an excellent choice.

However, it must be said that running a Web server is relatively easy, but running a well-managed server isn't. A poorly managed server is an open invitation to a world of problems, which can involve anywhere from invasion of privacy and breach of security to the server crawling to a halt or even crashing abruptly. We will discuss some of these issues in future work.

Acknowledgements

I would like to thank Paul Wrigglesworth, Department of Geography, Concordia University, Canada, for his assistance in testing Apache installation on Windows NT.

References

• Apache - Official site.
• Apache Documentation - Official Apache documentation.
• There are various Apache-specific books available. One of them is Apache: The Definitive Guide, Second Edition, By Ben Laurie & Peter Laurie, O'Reilly & Associates, 1999.
• There is an Apache quick reference card available in PostScript and PDF formats for various paper sizes, that can be quite handy at times.
• Apache Week - A free weekly electronic publication dedicated to the latest Apache developments. It provides an easy way to stay abreast of the latest news about Apache including: new releases, major bug fixes, plans for the next version, new apache ports and other related Apache information.
• The Apache FAQ
• comp.infosystems.www.servers.unix - UNIX variants, and comp.infosystems.www.servers.ms-windows for Windows 9x/NT.

Appendix : Apache Command Line Options

General

These options are applicable to all Apache ports, including UNIX and Windows 9x/NT.

Table 1. Apache General Command Line Options.

Windows-Specific

Table 2. Apache as a Console Application: Command Line Options.

Table 3. Apache as a Service: Command Line Options.

View the profile on Pankaj Kamthan and the list of other Articles by Pankaj Kamthan.