Server installation

The installation process follows the same basic pattern, whichever individual server you are installing:

Obtain the software, either in binary or in source form.
Compile the software if you obtained a source code distribution. If you obtained a binary distribution this step is irrelevant.
Move the software and ancillary files, such as configuration files, icons and support programs, into the appropriate locations and ensure that the file permissions are set correctly. Often there is an automated procedure for doing this.

Once you have installed the software you must configure it for your site. Configuration of Web servers is dealt with in Chapter .

Most servers are available by anonymous FTP as archive files which need to be unpacked. UNIX distributions are usually compressed tar files, which must be uncompressed before unpacking. Windows software is usually distributed as ZIP files. Before unpacking an archive file it is wise to examine the list of contents to determine whether the files will be extracted to a subdirectory or into the current directory. It is often safest to create a new directory and change to that directory to unpack the archive. This avoids overwriting existing files that have the same name, and mixing up files from the archive with existing directory contents.

When compiling software from a source code distribution it is sometimes necessary to make modifications in order to get the code to compile. Most Web servers have been ported to a number of different platforms, and are usually configured for a specific platform by editing a Makefile and/or a configuration header file (often called config.h). Instructions are usually given in a file called README, INSTALL or some such name, or in comments in the Makefile or header files. The settings of the CC and CFLAGS variables in the Makefile usually determine the compiler and compiler flags used. You may want to make sure that the maximum level of compiler optimization is specified, especially if you expect the server to be heavily loaded.

If the compiler complains that certain variables are undefined, especially if they have names in upper case, it is probable that a required system include file has been omitted. Use grep to search through the /usr/include and /usr/include/sys directories to see if the symbol is defined in one of the files there. If so try adding a `#include' statement to include that file in the source file that caused the errors, then recompile. If the compiler complains that it cannot find one of the include files, you may be able to use a version of the file from another system, but even if the compilation now proceeds with apparent success, you should test the server thoroughly as the replacement file may not be completely compatible. The copyright statement on some include files may prohibit you from using them in this way.

Another problem that sometimes occurs is that library functions that are not present on your system may be referenced. These are often functions standard on one version of UNIX but not on another. For example the function getwd on BSD and the getcwd function on System V UNIX both return the current working directory and can be implemented one in terms of the other. Many free software packages provide implementations of common library functions and it is often possible to extract the code from another package to supply the missing functions. You could use the Free Software Foundation's C library, glibc, which contains most common standard library functions, although if you want to distribute the server you are building, you should look at the licence and copyright statement.

Once the server has been compiled or if you have retrieved a binary distribution, the software components must be moved to appropriate directories. With commercial binary distributions this may be taken care of by an automated installation process, otherwise it is a matter of moving the files by hand or using the make utility if the Makefile includes an installation rule. Instructions in the distribution will normally indicate which files are needed and suggest appropriate locations for them.

Peculiarities of individual servers

Next: Peculiarities of individual Up: Web Software Installation Previous: Proxy and caching

[ITCP] Spinning the Web by Andrew Ford
© 1995 International Thomson Publishing
© 2002 Andrew Ford and Ford & Mason Ltd
Note: this HTML document was generated in December 1994 directly from the LaTeX source files using LaTeX2HTML. It was formatted into our standard page layout using the Template Toolkit. The document is mainly of historical interest as obviously many of the sites mentioned have long since disappeared.