Linux WWW HOWTO
 by Mr. Poet, [email protected]
 v0.85, 21 August 1999

 This document contains information about setting up WWW services under
 Linux (both server and client).  It tries not to be a in detail manual
 but an overview and a good pointer to further information.
 ______________________________________________________________________

 Table of Contents



 1. Introduction

    1.1 Copyright
    1.2 Feedback
    1.3 New versions of this Document

 2. Setting up WWW client software (Antiquated)

    2.1 Overview

 3. Lynx

    3.1 Where to get

 4. Emacs-W3

    4.1 Where to get

 5. Netscape Navigator/Communicator

    5.1 Different versions and options.
    5.2 Where to get
    5.3 Installing

 6. Setting up WWW server systems

    6.1 Overview

 7. Apache

    7.1 Where to get
    7.2 Compiling and Installing
    7.3 Configuring
    7.4 Hosting virtual websites
       7.4.1 IP based virtual hosting
       7.4.2 Shared IP virtual hosting
    7.5 CGI scripts
    7.6 Users Web Directories
    7.7 Daemon mode vs. Inetd mode
    7.8 Allowing put and delete commands
    7.9 User Authentication/Access Control
    7.10 su-exec
    7.11 Imagemaps
    7.12 SSI/XSSI
    7.13 Module system

 8. Web Server Add-ons

 9. Intranet Section

    9.1 What is required
    9.2 New versions of this document
    9.3 Feedback

 10. Install the HTTP server

    10.1 Preparation before downloading
       10.1.1 The Operating System
       10.1.2 Process type (ServerType)
       10.1.3 Binding Port (Port)
       10.1.4 Server user identity (User)
       10.1.5 Server group identity (Group)
       10.1.6 Server administrator email address (ServerAdmin)
       10.1.7 Location of server directory (ServerRoot)
       10.1.8 Location of HTML files (DocumentRoot)
    10.2 Compiling HTTPd
 11. Testing HTTPd

 12. Connecting to the Linux Server

    12.1 Setup the Linux server
    12.2 Setup the Netware server
    12.3 Setup the Netware Client
       12.3.1 Windows 3.x
       12.3.2 Windows 95
    12.4 Setup Microsoft Client
       12.4.1 Windows for Workgroups
       12.4.2 Windows 95
       12.4.3 Windows NT
    12.5 Setup TCP/IP on Macintosh
       12.5.1 MacTCP

 13. Setting up the Intranet

    13.1 NCPFS
       13.1.1 Installation
       13.1.2 Mounting NCPFS
    13.2 SMBFS
       13.2.1 Installation
       13.2.2 Mounting SMBFS
    13.3 NFS

 14. Accessing the Web

    14.1 Accessing the mounted filesystems
    14.2 Connecting to the Internet
    14.3 Other uses

 15. More things to do

 16. Credits and Legalities

    16.1 Thanks
    16.2 Copyright information

 17. For further reading

    17.1 O'Reilly & Associates Books
    17.2 Internet Request For Comments (RFC)


 ______________________________________________________________________

 1.  Introduction


 Many people are trying Linux because they are looking for a really
 good Internet capable operating system.  Also, there are institutes,
 universities, non-profits, and small businesses which want to set up
 Internet sites on a small budget. This is where the WWW-HOWTO comes
 in. This document explains how to set up clients and servers for the
 largest part of the Internet - The World Wide Web.

 All prices in this document are stated in US dollars.  This document
 assumes you are running Linux on an Intel platform.  Instructions and
 product availability my vary from platform to platform.  There are
 many links for downloading software in this document.  Whenever
 possible use a mirror site for faster downloading and to keep the load
 down on the main server.



 1.1.  Copyright

 This document is Copyright (c) 1999 by Mr. Poet

 This document is Copyright (c) 1997 by Wayne Leister. The original
 author of this document was Peter Dreuw.(All versions prior to 0.8)


 This HOWTO is free documentation; you can redistribute it and/or
 modify it under the terms of the GNU General Public License as
 published by the Free Software Foundation; either version 2 of the
 License, or (at your option) any later version.


 This document is distributed in the hope that it will be useful, but
 without any warranty; without even the implied warranty of
 merchantability or fitness for a particular purpose. See the GNU
 General Public License for more details.


 You can obtain a copy of the GNU General Public License by writing to
 the Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139,
 USA.

 Trademarks are owned by there respective owners.



 1.2.  Feedback

 Any feedback is welcome.  I do not claim to be an expert.  Some of
 this information was taken from badly written web sites; there are
 bound to be errors and omissions.  But make sure you have the latest
 version before you send corrections; It may be fixed in the next
 version (see the next section for where to get the latest version).
 Send feedback to [email protected].


 1.3.  New versions of this Document

 New versions of this document are always available at the LinuxPorts
 Website.



 2.  Setting up WWW client software (Antiquated)


      Please note that the following section is very outdated and
      was last updated in 1997. This section is being kept here
      for historical purposes only. All major distributions of
      Linux come with Netscape and Lynx. These are the two most
      popular browsers for the Linux Operating System. If you do
      not have these browsers installed currently please go to
      their respective sections for download information.


 The following chapter is dedicated to the setting up web browsers.
 Please feel free to contact me, if your favorite web browser is not
 mentioned here.   In this version of the document only a few of the
 browsers have there own section, but I tried to include all of them
 (all I could find) in the overview section.  In the future those
 browsers that deserve there own section will have it.
 The overview section is designed to help you decide which browser to
 use, and give you basic information on each browser.  The detail
 section is designed to help you install, configure, and maintain the
 browser.

 However I use Lynx when I don't feel like firing up the X-
 windows/Netscape monster.


 2.1.  Overview


    ``Navigator/Communicator''
       Netscape Navigator is the only Linux browser mentioned here,
       which is capable of advanced HTML features.  Some of these
       features are frames, Java, Javascript, automatic update, and
       layers.  It also has news and mail capability.  But it is a
       resource hog; it takes up lots of CPU time and memory.  It also
       sets up a separate cache for each user wasting disk space.

       Netscape is currently an OpenSource product and can be
       downloaded from ftp.netscape.com=>.  <tag><ref id= name="Lynx">
       Lynx is the one of the smallest web browsers.  It is the king of
       text based browsers.  It's free and the source code is available
       under the GNU public license.  It's text based, but it has many
       special features. Lynx now supports tables, color (via curses)
       and frames.


         Note on frame support for lynx:


    The frame support for lynx is limited, it will notice the frames
    and show the title of the frames for your to select as hot links.
    Since, frame titles are usually very undescriptive for coding
    simplicity this can be confusing.



    Kfm
       Kfm is part of the K Desktop Environment (KDE).  KDE is a system
       that runs on top of X-windows.  It gives you many features like
       drag an drop, sounds, a trashcan and a unified look and feel.
       Kfm is the K File Manager, but it is also a web browser. It is
       very usable as a web browser and it supports frames, tables, ftp
       downloads, looking into tar files, and more.  The current
       release of KDE is 1.1.1 with 1.1.2 very near by.  Kfm can be
       used without KDE, but you still need the librarys that come with
       KDE.  For more information about KDE and Kfm visit the KDE
       website at  <http://www.kde.org>.


    ``Emacs''
       Emacs is the one program that does everything.  It is a word
       processor, news reader, mail reader, and web browser.  It has a
       steep learning curve at first, because you have to learn what
       all the keys do.  The X-windows version is easier to use,
       because most of the functions are on menus.  Another drawback is
       that it's mostly text based. (It can display graphics if you are
       running it under X-windows).  It is also free, and the source
       code is available under the GNU public license.


    NCSA Mosaic
       Mosaic is an X-windows browser developed by the National Center
       for Supercomputing Applications (NCSA) at the University of
       Illinois.  NCSA spent four years on the project and has now
       moved on to other things.  Again, Mosaic is no longer supported.
       However since the source is free for non-commercial use it might
       make an interesting project for someone who wants to develop a
       new browser.


    Amaya
       Amaya is the X-windows concept browser for the W3C for HTML 3.2.
       Therefore it supports all the HTML 3.2 standards.  It also
       supports some of the features of HTML 4.0.  It supports tables,
       forms, client side image maps, put publishing, gifs, jpegs, and
       png graphics.  It is both a browser and authoring tool.  The
       latest public release is 1.0 beta.  Version 1.1 beta is in
       internal testing and is due out soon.  For more information
       visit the Amaya web site at  <http://www.w3.org/Amaya/>.  It can
       be downloaded from <ftp://ftp.w3.org/pub/Amaya-LINUX-
       ELF-1.0b.tar.gz>.


    Qweb
       Qweb is yet another basic X-windows browser.  It supports
       tables, forms, and server site image maps.  The latest version
       is 1.3.  For more information visit the Qweb website at
       <http://sunsite.auc.dk/qweb/> The source is available from
       <http://sunsite.auc.dk/qweb/qweb-1.3.tar.gz> The binaries are
       available in a Red Hat RPM from
       <http://sunsite.auc.dk/qweb/qweb-1.3-1.i386.rpm>


 It is the reccomendation of this author that users of web browsers use
 either Netscape 4.x, Lynx or Netscape 5.xAlpha. They are the only one
 currently available for Linux that support most features. Personally I
 suggest Netscape5.xAlpha, even though it is Alpha Software it is quite
 stable and frankly implements the standards better than the 4.x
 versions do.



 3.  Lynx

 Lynx is one of the smaller (around 600 K executable) and faster web
 browsers available. It does not eat up much bandwidth nor system
 resources as it only deals with text displays.  It can display on any
 console, terminal or xterm. You will not need an  X Windows system or
 additional system memory to run this little browser.


 3.1.  Where to get

 Most distributions have Lynx in them.  Therefore I will not bore you
 with the details of compiling and installing Lynx.

 The latest version is 2.8.2 and can be retrieved from
 <http://www.slcc.edu/lynx/fote/> or from almost any friendly Linux FTP
 server like ftp://sunsite.unc.edu under /pub/Linux/apps/www/broswers/
 or mirror site.

 For more information on Lynx try these locations:

    Lynx Links
       <http://www.crl.com/~subir/lynx.html>


    Lynx Pages
       <http://lynx.browser.org>

    Lynx Help Pages
       <http://www.crl.com/~subir/lynx/lynx_help/lynx_help_main.html>
       (the same pages you get from lynx --help and typing ? in lynx)

 Note: The Lynx help pages have recently moved.  If you have an older
 version of Lynx, you will need to change your lynx.cfg (in /usr/lib)
 to point to the new address(above).

 I think the most special feature of Lynx against all other web
 browsers is the capability for batch mode retrieval. One can write a
 shell script which retrieves a document, file or anything like that
 via http, FTP, gopher, WAIS, NNTP or file:// - url's and save it to
 disk.  Furthermore, one can fill in data into HTML forms in batch mode
 by simply redirecting the standard input and using the -post_data
 option.

 For more special features of Lynx just look at the help files and the
 man pages.  If you use a special feature of Lynx that you would like
 to see added to this document, let me know.



 4.  Emacs-W3

 There are several different flavors of Emacs.  The two most popular
 are GNU Emacs and XEmacs.  GNU Emacs is put out by the Free Software
 Foundation, and is the original Emacs.  It is mainly geared toward
 text based terminals, but it does run in X-Windows.  XEmacs (formerly
 Lucid Emacs) is a version that only runs on X-Windows.  It has many
 special features that are X-Windows related (better menus etc).


 4.1.  Where to get


 Most distributions include GNU Emacs.

 The most recent GNU emacs is 19.34.  It doesn't seem to have a web
 site.  The FTP site is at  <ftp://ftp.gnu.ai.mit.edu/pub/gnu/>.

 The latest version of XEmacs is 20.2.  The XEmacs FTP site is at
 <ftp://ftp.xemacs.org/pub/xemacs>.  For more information about XEmacs
 goto see its web page at <http://www.xemacs.org>.

 Both are available from the Linux archives at ftp://sunsite.unc.edu
 under /pub/Linux/apps/editors/emacs/


 If you got GNU Emacs or XEmacs installed, you probably got the W3
 browser running to.

 The Emacs W3 mode is a nearly fully featured web browser system
 written in the Emacs Lisp system. It mostly deals with text, but can
 display graphics, too - at least - if you run the emacs under the X
 Window system.

 To get XEmacs in to W3 mode, goto the apps menu and select browse the
 web.

 I don't use Emacs, so if someone will explain how to get it into the
 W3 mode I'll add it to this document.  Most of this information was
 from the original author.  If any information is incorrect, please let
 me know.  Also let me know if you think anything else should be added
 about Emacs.



 5.  Netscape Navigator/Communicator


 5.1.  Different versions and options.

 Netscape Navigator is the King of WWW browsers.  Netscape Navigator
 can do almost everything. But on the other hand, it is one of the most
 memory hungry and resource eating program I've ever seen.

 There are 3 different versions of the program:

 Netscape Navigator includes the web browser, netcaster (push client)
 and a basic mail program.

 Netscape Communicator includes the web browser, a web editor, an
 advanced mail program, a news reader, netcaster (push client), and a
 group conference utility.

 Netscape Communicator Pro includes everything Communicator has plus a
 group calendar, IBM terminal emulation, and remote administration
 features (administrators can update thousands of copies of Netscape
 from their desk).

 In addition to the three versions there are two other options you must
 pick.

 The first is full install or base install.  The full install includes
 everything.  The base install includes enough to get you started.  You
 can download the additional components as you need them (such as
 multimedia support and netcaster).  These components can be installed
 by the Netscape smart update utility (after installing goto
 help->software updates).  At this time the full install is not
 available for Linux.

 The second option is import or export.  If you are from the US are
 Canada you have the option of selecting the import version.  This
 gives you the stronger 128 bit encryption for secure transactions
 (SSL).  The export version only has 40 bit encryption, and is the only
 version allowed outside the US and Canada.

 The latest version of the Netscape Navigator/Communicator/Communicator
 Pro is 4.6.1.  There are two different versions for Linux.  One is for
 the old 1.2 series kernels and one for the new 2.x kernels.  If you
 don't have at least a 2.0 kernel I suggest you upgrade; there are many
 improvements in the new kernel.

 Beta versions are also available.  If you try a beta version, they
 usually expire in a month or so!

 You can also try the 5.0 alpha at the Mozilla project located at
 www.mozilla.org. I suggest this highly, the new Netscape 5 is the best
 browser I have seen in a long time. It is extremely flexible and
 mostly stable even for an Alpha release.

 5.2.  Where to get


 The best way to get Netscape software is to go through their web site
 at <http://www.netscape.com/download/>.  They have menu's to guide you
 through the selection.  When it ask for the Linux version, it is
 referring to the kernel (most people should be using 2.0 by now).  If
 your not sure which version kernel you have run 'cat /proc/version'.
 Going through the web site is the only way to get the import versions.

 If you want an export version you can download them directly from the
 Netscape FTP servers.  The FTP servers are also more up to date. For
 example when I first wrote this the web interface did not have the
 non-beta 4.03 for Linux yet, but it was on the FTP site.  Here are the
 links to the export Linux 2.0 and 2.2 versions:

 Netscape Navigator 4.6.1 is at
 <ftp://ftp.netscape.com/pub/communicator/4.6/shipping/english/unix/linux20/navigator_standalone/navigator-
 v403-export.x86-unknown-linux2.0.tar.gz>

 Netscape Communicator 4.6.1 for Linux 2.0 (kernel) is at
 <ftp://ftp.netscape.com/pub/communicator/4.6/shipping/english/unix/linux20/base_install/communicator-
 v403-export.x86-unknown-linux2.0.tar.gz>

 These url's will change as new versions come out.  If these links
 break you can find them by fishing around at the FTP site
 <ftp://ftp.netscape.com/pub/communicator/>.

 These servers are heavily loaded at times.  Its best to wait for off
 peak hours or select a mirror site.  Be prepared to wait, these
 archives are large.  Navigator is almost 8megs, and Communicator base
 install is 10megs.


 5.3.  Installing

 This section explains how to install version 4 of Netscape Navigator,
 Communicator, and Communicator Pro.

 First unpack the archive to a temporary directory.  Then run the ns-
 install script (type ./ns-install).  Then make a symbolic link from
 the /usr/local/netscape/netscape binary to /usr/local/bin/netscape
 (type ln -s /usr/local/netscape/netscape /usr/local/bin/netscape).
 Finally set the system wide environment variable $MOZILLA_HOME to
 /usr/local/netscape so Netscape can find its files.  If you are using
 bash for your shell edit your /etc/profile and add the lines:



      MOZILLA_HOME="/usr/local/netscape"
      export MOZILLA_HOME



 After you have it installed the software can automatically update
 itself with smart update.  Just run Netscape as root and goto
 help->software updates.  If you only got the base install, you can
 also install the Netscape components from there.

 Note: This will not remove any old versions of Netscape, you must
 manually remove them by deleting the Netscape binary and Java class
 file (for version 3).



 6.  Setting up WWW server systems

 This section contains information on different http server software
 packages and additional server side tools like script languages for
 CGI programs etc.  There are several dozen web servers, I only covered
 those that are fully functional.  As some of these are commercial
 products, I have no way of trying them.  Most of the information in
 the overview section was pieced together from various web sites.  If
 there is any incorrect or missing information please let me know.

 For a technical description on the http mechanism, take a look at the
 RFC documents mentioned in the chapter "For further reading" of this
 HOWTO.

 I prefer to use the Apache server.  It has almost all the features you
 would ever need and its free!  I will admit that this section is
 heavily biased toward Apache.  I decided to concentrate my efforts on
 the Apache section rather than spread it out over all the web servers.
 I may cover other web servers in the future.



 6.1.  Overview


    Cern httpd
       This was the first web server.  It was developed by the European
       Laboratory for Particle Physics (CERN).  CERN httpd is no longer
       supported.  The CERN httpd server is reported to have some ugly
       bugs, to be quite slow and resource hungry.  The latest version
       is 3.0.  For more information visit the CERN httpd home page at
       <http://www.w3.org/Daemon/Status.html>.  It is available for
       download at
       <ftp://sunsite.unc.edu/pub/Linux/apps/www/servers/httpd-3.0.term.tpz>
       (no it is not a typo, the extension is actually .tpz on the
       site; probably should be .tgz)


    NCSA HTTPd
       The NCSA HTTPd server is the father to Apache (The development
       split into two different servers).   Therefore the setup files
       are very similar.  NCSA HTTPd is free and the source code is
       available.  This server not covered in this document, although
       reading the Apache section may give you some help.  The NCSA
       server was once popular, but most people are replacing it with
       Apache.  Apache is a drop in replacement for the NCSA
       server(same configuration files), and it fixes several
       shortcomings of the NCSA server.  NCSA HTTPd accounts for 4.9%
       (and falling) of all web servers. (source September 1997
       Netcraft survey <http://www.netcraft.com/survey/>).  The latest
       version is 1.5.2a.  For more information see the NCSA website at
       <http://hoohoo.ncsa.uiuc.edu>.


    ``Apache''
       Apache is the king of all web servers.  Apache and its source
       code is free.  Apache is modular, therefore it is easy to add
       features.  Apache is very flexible and has many, many features.
       Apache and its derivatives makes up 55% of all web domains.
       There are over 3,928,112 Apache servers in operation (source
       August 1999 Netcraft survey <http://www.netcraft.com/survey/>).

       The official Apache is missing SSL, but there are two
       derivatives that fill the gap.  Stronghold is a commercial
       product that is based on Apache.  It retails for $995; an
       economy version is available for $495 (based on an old version
       of Apache).  Stronghold is the number two secure server behind
       Netscape (source C2 net <http://www.c2.net/products/stronghold>
       and Netcraft survey <http://www.netcraft.com/survey/>).  For
       more information visit the Stronghold website at
       <http://www.c2.net/products/stronghold/>.  It was developed
       outside the US, so it is available with 128 bit SSL everywhere.

       Apache-SSL is a free implementation of SSL, but it is not for
       commercial use in the US (RSA has US patents on SSL technology).
       It can be used for non-commercial use in the US if you link with
       the free RSAREF library.  For more information see the website
       at <http://www.algroup.co.uk/Apache-SSL/>.


    Netscape Fast Track Server
       Fast Track was developed by Netscape, but the Linux version is
       put out by Caldera.  The Caldera site lists it as Fast Track for
       OpenLinux.  I'm not sure if it only runs on Caldera OpenLinux or
       if any Linux distribution will do (E-mail me if you have the
       answer).  Netscape servers account for 6.11% (and RISING!) of
       all web servers (source August 1999
       <http://www.netcraft.com/survey/>).  The server sells for $295.
       It is also included with the Caldera OpenLinux Standard
       distribution which sells for $399 ($199.50 educational).  The
       web pages tell of a nice administration interface and a quick 10
       minute setup.  The server has support for 40-bit SSL.  To get
       the full 128-bit SSL you need Netscape Enterprise Server.
       Unfortunately that is not available for Linux :( The latest
       version available for Linux is 2.0 (Version 3 is in beta, but
       its not available for Linux yet).  To buy a copy goto the
       Caldera web site at
       <http://www.caldera.com/products/netscape/netscape.html> For
       more information goto the Fast Track page at
       <http://www.netscape.com/comprod/server_central/product/fast_track/>



    WN WN has many features that make it attractive.  First it is
       smaller than the CERN, NCSA HTTPd, an Apache servers.  It also
       has many built-in features that would require CGI's.  For
       example site searches, enhanced server side includes.  It can
       also decompress/compress files on the fly with its filter
       feature.  It also has the ability to retrieve only part of a
       file with its ranges feature.  It is released under the GNU
       public license.  The current version is 2.08. For more
       information see the WN website at <http://hopf.math.nwu.edu/>.


    AOLserver
       AOLserver is made by America Online.  I'll admit that I was
       surprised by the features of a web server coming from AOL.  In
       addition to the standard features it supports database
       connectivity.  Pages can query a database by Structured Query
       Language (SQL) commands.  The database is access through Open
       Database Connectivity (ODBC).  It also has built-in search
       engine and TCL scripting.  If that is not enough you can add
       your own modules through the c Application Programming Interface
       (API).  I almost forgot to mention support for 40 bit SSL.  And
       you get all this for free!  For more information visit the
       AOLserver site at <http://www.aolserver.com/server/>


    CL-HTTP
       CL-HTTP stands for Common Lisp Hypermedia Server.  If you are a
       Lisp programmer this server is for you.  You can write your CGI
       scripts in Lisp.  It has a web based setup function.  It also
       supports all the standard server features.  CL-HTTP is free and
       the source code is available.  For more information visit the
       CL-HTTP website at <http://www.ai.mit.edu/projects/iiip/doc/cl-
       http/home-page.html>

 If you have a commercial purpose (company web site, or ISP), I would
 strongly recommend that you use Apache.  I've also heard that the
 Netscape Server is easy to setup. If you have an internal use you can
 be a bit more flexible. But unless one of them has a feature that you
 just have to use,

 This is only a partial listing of all the servers available.  For a
 more complete list visit Netcraft at
 <http://www.netcraft.com/survey/servers.html> or Web Compare at
 <http://webcompare.internet.com>.



 7.  Apache

 The current version of Apache is 1.3.9.  The main Apache site is at
 <http://www.apache.org/>.  Another good source of information is
 Apacheweek at <http://www.apacheweek.com/>.  The Apache documentation
 is ok, so I'm not going to go into detail in setting up apache.  The
 documentation is on the website and is included with the source (in
 HTML format).  There are also text files included with the source, but
 the HTML version is better.  The documentation should get a whole lot
 better once the Apache Documentation Project gets under way.  Right
 now most of the documents are written by the developers.  Not to
 discredit the developers, but they are a little hard to understand if
 you don't know the terminology.


 7.1.  Where to get

 Apache is included in the Red Hat, Slackware, and OpenLinux
 distributions.  Although they may not be the latest version, they are
 very reliable binaries.  The bad news is you will have to live with
 their directory choices (which are totally different from each other
 and the Apache defaults).

 The source is available from the Apache web site at
 <http://www.apache.org/dist/> Binaries are are also available at
 apache at the same place.  You can also get binaries from sunsite at
 <ftp://sunsite.unc.edu/pub/Linux/apps/www/servers/>.  And for those of
 us running Red Hat the latest binary RPM file can usually be found in
 the contrib directory at <ftp://ftp.redhat.com/pub/contrib/i386/>

 If your server is going to be used for commercial purposes, it is
 highly recommended that you get the source from the Apache website and
 compile it yourself.  The other option is to use a binary that comes
 with a major distribution.  For example Slackware, Red Hat, or
 OpenLinux distributions.  The main reason for this is security.  An
 unknown binary could have a back door for hackers, or an unstable
 patch that could crash your system. This also gives you more control
 over what modules are compiled in, and allows you to set the default
 directories.  It's not that difficult to compile Apache, and besides
 you not a real Linux user until you compile your own programs ;)



 7.2.  Compiling and Installing

 First untar the archive to a temporary directory.  Next change to the
 src directory.  Then edit the Configuration file if you want to
 include any special modules.  The most commonly used modules are
 already included.  There is no need to change the rules or makefile
 stuff for Linux.  Next run the Configure shell script (./Configure).
 Make sure it says Linux platform and gcc as the compiler.  Next you
 may want to edit the httpd.h file to change the default directories.
 The server home (where the config files are kept) default is
 /usr/local/etc/httpd/, but you may want to change it to just
 /etc/httpd/.  And the server root (where the HTML pages are served
 from) default is /usr/local/etc/httpd/htdocs/, but I like the
 directory /home/httpd/html (the Red Hat default for Apache).  If you
 are going to be using su-exec (see special features below) you may
 want to change that directory too.  The server root can also be
 changed from the config files too.  But it is also good to compile it
 in, just encase Apache can't find or read the config file.  Everything
 else should be changed from the config files.  Finally run make to
 compile Apache.

 If you run in to problems with include files missing, check the
 following things. Make sure you have the kernel headers (include
 files) installed for your kernel version. Also make sure you have
 these symbolic links in place:


      /usr/include/linux should be a link to /usr/src/linux/include/linux
      /usr/include/asm should be a link to /usr/src/linux/include/asm
      /usr/src/linux should be a link to the Linux source directory (ex.linux-2.0.30)



 Links can be made with ln -s, it works just like the cp command except
 it makes a link (ln -s source-dir destination-link)

 When make is finished there should be an executable named httpd in the
 directory.  This needs to be moved in to a bin directory.  /usr/sbin
 or /usr/local/sbin would be good choices.

 Copy the conf, logs, and icons sub-directories from the source to the
 server home directory.  Next rename 3 of the files files in the conf
 sub-directory to get rid of the -dist extension (ex. httpd.conf-dist
 becomes httpd.conf)

 There are also several support programs that are included with Apache.
 They are in the support directory and must be compiled and installed
 separately.  Most of them can be make by using the makefile in that
 directory (which is made when you run the main Configure script).  You
 don't need any of them to run Apache, but some of them make the
 administrators job easier.


 7.3.  Configuring

 Now you should have four files in your conf sub-directory (under your
 server home directory).  The httpd.conf sets up the server daemon
 (port number, user, etc).  The srm.conf sets the root document tree,
 special handlers, etc.  The access.conf sets the base case for access.
 Finally mime.types tells the server what mime type to send to the
 browser for each extension.

 The configuration files are pretty much self-documented (plenty of
 comments), as long as you understand the lingo.  You should read
 through them thoroughly before putting your server to work.  Each
 configuration item is covered in the Apache documentation.

 The mime.types file is not really a configuration file.  It is used by
 the server to translate file extensions into mime-types to send to the
 browser.  Most of the common mime-types are already in the file.  Most
 people should not need to edit this file.  As time goes on, more mime
 types will be added to support new programs.  The best thing to do is
 get a new mime-types file (and maybe a new version of the server) at
 that time.

 Always remember when you change the configuration files you need to
 restart Apache or send it the SIGHUP signal with kill for the changes
 to take effect.  Make sure you send the signal to the parent process
 and not any of the child processes.  The parent usually has the lowest
 process id number.  The process id of the parent is also in the
 httpd.pid file in the log directory.  If you accidently send it to one
 of the child processes the child will die and the parent will restart
 it.

 I will not be walking you through the steps of configuring Apache.
 Instead I will deal with specific issues, choices to be made, and
 special features.

 I highly recommend that all users read through the security tips in
 the Apache documentation.  It is also available from the Apache
 website at <http://www.apache.org/docs/mics/security_tips.html>.


 7.4.  Hosting virtual websites

 Virtual Hosting is when one computer has more than one domain name.
 The old way was to have each virtual host have its own IP address.
 The new way uses only one IP address, but it doesn't work correctly
 with browsers that don't support HTTP 1.1.

 My recommendation for businesses is to go with the IP based virtual
 hosting until most people have browsers that support HTTP 1.1 (give it
 a year or two).   This also gives you a more complete illusion of
 virtual hosting.  While both methods can give you virtual mail
 capabilities (can someone confirm this?), only IP based virtual
 hosting can also give you virtual FTP as well.

 If it is for a club or personal page, you may want to consider shared
 IP virtual hosting.  It should be cheaper than IP based hosting and
 you will be saving precious IP addresses.

 You can also mix and match IP and shared IP virtual hosts on the same
 server.  For more information on virtual hosting visit Apacheweek at
 <http://www.apacheweek.com/features/vhost>.


 7.4.1.  IP based virtual hosting

 In this method each virtual host has its own IP address.  By
 determining the IP address that the request was sent to, Apache and
 other programs can tell what domain to serve.  This is an incredible
 waste of IP space.  Take for example the servers where my virtual
 domain is kept.  They have over 35,000 virtual accounts, that means
 35,000 IP addresses.  Yet I believe at last count they had less than
 50 servers running.

 Setting this up is a two part process.  The first is getting Linux
 setup to accept more than one IP address.  The second is setting up
 apache to serve the virtual hosts.

 The first step in setting up Linux to accept multiple IP addresses is
 to make a new kernel.  This works best with a 2.0 series kernel (or
 higher).  You need to include IP networking and IP aliasing support.
 If you need help with compiling the kernel see the kernel howto
 <http://sunsite.unc.edu/LDP/HOWTO/Kernel-HOWTO.html>.

 Next you need to setup each interface at boot.  If you are using the
 Red Hat Distribution then this can be done from the control panel.
 Start X-windows as root, you should see a control panel.  Then double
 click on network configuration.  Next goto the interfaces panel and
 select your network card.  Then click alias at the bottom of the
 screen.  Fill in the information and click done.  This will need to be
 done for each virtual host/IP address.

 If you are using other distributions you may have to do it manually.
 You can just put the commands in the rc.local file in /etc/rc.d
 (really they should go in with the networking stuff).  You need to
 have a ifconfig and route command for each device.  The aliased
 addresses are given a sub device of the main one.  For example eth0
 would have aliases eth0:0, eth0:1, eth0:2, etc.  Here is an example of
 configuring a aliased device:


      ifconfig eth0:0 192.168.1.57
      route add -host 192.168.1.57 dev eth0:0



 You can also add a broadcast address and a netmask to the ifconfig
 command.  If you have alot of aliases you may want to make a for loop
 to make it easier.  For more information see the IP alias mini howto
 <http://sunsite.unc.edu/LDP/HOWTO/mini/IP-Alias.html>.

 Then you need to setup your domain name server (DNS) to serve these
 new domains.   And if you don't already own the domain names, you need
 to contact the Internic <http://www.internic.net> to register the
 domain names.  See the DNS-howto for information on setting up your
 DNS.

 Finally you need to setup Apache to server the virtual domain
 correctly.  This is in the httpd.conf configuration file near the end.
 They give you an example to go by.  All commands specific to that
 virtual host are put in between the virtualhost directive tags.  You
 can put almost any command in there.  Usually you set up a different
 document root, script directory, and log files. You can have almost
 unlimited number of virtual hosts by adding more virtualhost directive
 tags.

 In rare cases you may need to run separate servers if a directive is
 needed for a virtual host, but is not allowed in the virtual host
 tags.  This is done using the bindaddress directive.   Each server
 will have a different name and setup files.  Each server only responds
 to one IP address, specified by the bindaddress directive.  This is an
 incredible waste of system resources.


 7.4.2.  Shared IP virtual hosting

 This is a new way to do virtual hosting.  It uses a single IP address,
 thus conserving IP addresses for real machines (not virtual ones).  In
 the same example used above those 30,000 virtual hosts would only take
 50 IP addresses (one for each machine).  This is done by using the new
 HTTP 1.1 protocol.  The browser tells the server which site it wants
 when it sends the request.  The problem is browsers that don't support
 HTTP 1.1 will get the servers main page, which could be setup to
 provide a menu of virtual hosts available.  That ruins the whole
 illusion of virtual hosting.  The illusion that you have your own
 server.

 The setup is much simpler than the IP based virtual hosting.  You
 still need to get your domain from the Internic and setup your DNS.
 This time the DNS points to the same IP address as the original
 domain.  Then Apache is setup the same as before.  Since you are using
 the same IP address in the virtualhost tags, it knows you want Shared
 IP virtual hosting.

 There are several work arounds for older browsers.  I'll explain the
 best one.  First you need to make your main pages a virtual host
 (either IP based or shared IP).  This frees up the main page for a
 link list to all your virtual hosts.  Next you need to make a back
 door for the old browsers to get in.  This is done using the
 ServerPath directive for each virtual host inside the virtualhost
 directive.  For example by adding ServerPath /mysite/ to
 www.mysite.com old browsers would be able to access the site by
 www.mysite.com/mysite/.  Then you put the default page on the main
 server that politely tells them to get a new browser, and lists links
 to all the back doors of all the sites you host on that machine.  When
 an old browser accesses the site they will be sent to the main page,
 and get a link to the correct page.  New browsers will never see the
 main page and will go directly to the virtual hosts.  You must
 remember to keep all of your links relative within the web sites,
 because the pages will be accessed from two different URL's
 (www.mysite.com and www.mysite.com/mysite/).

 I hope I didn't lose you there, but its not an easy workaround.  Maybe
 you should consider IP based hosting after all.  A very similar
 workaround is also explained on the apache website at
 <http://www.apache.org/manual/host.html>.

 If anyone has a great resource for Shared IP hosting, I would like to
 know about it.  It would be nice to know what percent of browsers out
 there support HTTP 1.1, and to have a list of which browsers and
 versions support HTTP 1.1.


 7.5.  CGI scripts

 There are two different ways to give your users CGI script capability.
 The first is make everything ending in .cgi a CGI script.  The second
 is to make script directories (usually named cgi-bin).  You could also
 use both methods.  For either method to work the scripts must be world
 executable (chmod 711).  By giving your users script access you are
 creating a big security risk.  Be sure to do your homework to minimize
 the security risk.

 I prefer the first method, especially for complex scripting.  It
 allows you to put scripts in any directory.  I like to put my scripts
 with the web pages they work with.  For sites with allot of scripts it
 looks much better than having a directory full of scripts.  This is
 simple to setup.  First uncomment the .cgi handler at the end of the
 srm.conf file.  Then make sure all your directories have the option
 ExecCGI or All in the access.conf file.

 Making script directories is considered more secure.  To make a script
 directory you use the ScriptAlias directive in the srm.conf file.  The
 first argument is the Alias the second is the actual directory.  For
 example ScriptAlias /cgi-bin/ /usr/httpd/cgi-bin/ would make
 /usr/httpd/cgi-bin able to execute scripts.  That directory would be
 used whenever someone asked for the directory /cgi-bin/.  For security
 reasons you should also change the properties of the directory to
 Options none, AllowOveride none in the access.conf (just uncomment the
 example that is there).  Also do not make your script directories
 subdirectories of your web page directories.  For example if you are
 serving pages from /home/httpd/html/, don't make the script directory
 /home/httpd/html/cgi-bin; Instead make it /home/httpd/cgi-bin.

 If you want your users to have there own script directories you can
 use multiple ScriptAlias commands.  Virtual hosts should have there
 ScriptAlias command inside the virtualhost directive tags.  Does
 anyone know a simple way to allow all users to have a cgi-bin
 directory without individual ScriptAlias commands?



 7.6.  Users Web Directories

 There are two different ways to handle user web directories.  The
 first is to have a subdirectory under the users home directory
 (usually public_html).  The second is to have an entirely different
 directory tree for web directories.  With both methods make sure set
 the access options for these directories in the access.conf file.

 The first method is already setup in apache by default.  Whenever a
 request for /~bob/ comes in it looks for the public_html directory in
 bob's home directory.  You can change the directory with the UserDir
 directive in the srm.conf file.  This directory must be world readable
 and executable.  This method creates a security risk because for
 Apache to access the directory the users home directory must be world
 executable.

 The second method is easy to setup.  You just need to change the
 UserDir directive in the srm.conf file.  It has many different
 formats; you may want to consult the Apache documentation for
 clarification.  If you want each user to have their own directory
 under /home/httpd/, you would use UserDir /home/httpd.  Then when the
 request /~bob/ comes in it would translate to /home/httpd/bob/.  Or if
 you want to have a subdirectory under bob's directory you would use
 UserDir /home/httpd/*/html.  This would translate to
 /home/httpd/bob/html/ and would allow you to have a script directory
 too (for example /home/httpd/bob/cgi-bin/).


 7.7.  Daemon mode vs. Inetd mode

 There are two ways that apache can be run.  One is as a daemon that is
 always running (Apache calls this standalone).  The second is from the
 inetd super-server.

 Daemon mode is far superior to inetd mode.  Apache is setup for daemon
 mode by default.  The only reason to use the inetd mode is for very
 low use applications.  Such as internal testing of scripts, small
 company Intranet, etc.  Inetd mode will save memory because apache
 will be loaded as needed.   Only the inetd daemon will remain in
 memory.

 If you don't use apache that often you may just want to keep it in
 daemon mode and just start it when you need it.  Then you can kill it
 when you are done (be sure to kill the parent and not one of the child
 processes).

 To setup inetd mode you need to edit a few files.  First in
 /etc/services see if http is already in there.  If its not then add
 it:


      http    80/tcp



 Right after 79 (finger) would be a good place.  Then you need to edit
 the /etc/inetd.conf file and add the line for Apache:



 http    stream  tcp     nowait  root    /usr/sbin/httpd httpd



 Be sure to change the path if you have Apache in a different location.
 And the second httpd is not a typo; the inet daemon requires that.  If
 you are not currently using the inet daemon, you may want to comment
 out the rest of the lines in the file so you don't activate other ser-
 vices as well (FTP, finger, telnet, and many other things are usually
 run from this daemon).

 If you are already running the inet deamon (inetd), then you only need
 to send it the SIGHUP signal (via kill; see kill's man page for more
 info) or reboot the computer for changes to take effect.  If you are
 not running inetd then you can start it manually.  You should also add
 it to your init files so it is loaded at boot (the rc.local file may
 be a good choice).


 7.8.  Allowing put and delete commands

 The newer web publishing tools support this new method of uploading
 web pages by http (instead of FTP).  Some of these products don't even
 support FTP anymore!  Apache does support this, but it is lacking a
 script to handle the requests.  This script could be a big security
 hole, be sure you know what you are doing before attempting to write
 or install one.

 If anyone knows of a script that works let me know and I'll include
 the address to it here.

 For more information goto Apacheweek's article at
 <http://www.apacheweek.com/features/put>.


 7.9.  User Authentication/Access Control


 This is one of my favorite features.  It allows you to password
 protect a directory or a file without using CGI scripts.  It also
 allows you to deny or grant access based on the IP address or domain
 name of the client.  That is a great feature for keeping jerks out of
 your message boards and guest books (you get the IP or domain name
 from the log files).

 To allow user authentication the directory must have AllowOverrides
 AuthConfig set in the access.conf file.  To allow access control (by
 domain or IP address) AllowOverrides Limit must be set for that
 directory.

 Setting up the directory involves putting an .htaccess file in the
 directory.  For user authentication it is usually used with an
 .htpasswd and optionally a .htgroup file.  Those files can be shared
 among multiple .htaccess files if you wish.

 For security reasons I recommend that everyone use these directives in
 there access.conf file:



      <files ~ "/\.ht">
      order deny,allow
      deny from all
      </files>

 If you are not the administrator of the system you can also put it in
 your .htaccess file if AllowOverride Limit is set for your directory.
 This directive will prevent people from looking into your access
 control files (.htaccess, .htpasswd, etc).

 There are many different options and file types that can be used with
 access control.  Therefore it is beyond the scope of this document to
 describe the files.  For information on how to setup User
 Authentication see the Apacheweek feature at
 <http://www.apacheweek.com/features/userauth> or the NCSA pages at
 <http://hoohoo.ncsa.uiuc.edu/docs-1.5/tutorials/user.html>.


 7.10.  su-exec

 The su-exec feature runs CGI scripts as the user of the owner.
 Normally it is run as the user of the web server (usually nobody).
 This allows users to access there own files in CGI scripts without
 making them world writable (a security hole).  But if you are not
 careful you can create a bigger security hole by using the su-exec
 code.  The su-exec code does security checks before executing the
 scripts, but if you set it up wrong you will have a security hole.

 The su-exec code is not for amateurs.  Don't use it if you don't know
 what you are doing.  You could end up with a gaping security hole
 where your users can gain root access to your system.  Do not modify
 the code for any reason.  Be sure to read all the documentation
 carefully.  The su-exec code is hard to setup on purpose, to keep the
 amateurs out (everything must be done manually, no make file no
 install scripts).

 The su-exec code resides in the support directory of the source.
 First you need to edit the suexec.h file for your system.  Then you
 need to compile the su-exec code with this command:


      gcc suexec.c -o suexec



 Then copy the suexec executable to the proper directory.  The Apache
 default is /usr/local/etc/httpd/sbin/.  This can be changed by editing
 httpd.h in the Apache source and recompiling Apache.  Apache will only
 look in this directory, it will not search the path.  Next the file
 needs to be changed to user root (chown root suexec) and the suid bit
 needs to be set (chmod 4711 suexec).  Finally restart Apache, it
 should display a message on the console that su-exec is being used.

 CGI scripts should be set world executable like normal.  They will
 automaticaly be run as the owner of the CGI script.  If you set the
 SUID (set user id) bit on the CGI scripts they will not run.  If the
 directory or file is world or group writable the script will not run.
 Scripts owned by system users will not be run (root, bin, etc.).  For
 other security conditions that must be met see the su-exec
 documentation.  If you are having problems see the su-exec log file
 named cgi.log.

 Su-exec does not work if you are running Apache from inetd, it only
 works in daemon mode.  It will be fixed in the next version because
 there will be no inetd mode.  If you like playing around in source
 code, you can edit the http_main.c.  You want to get rid of the line
 where Apache announces that it is using the su-exec wrapper (It
 wrongly prints this in front of the output of everything).


 Be sure and read the Apache documentation on su-exec.  It is included
 with the source and is available on the Apache web site at
 <http://www.apache.org/docs/suexec.html>


 7.11.  Imagemaps

 Apache has the ability to handle server side imagemaps.  Imagemaps are
 images on webpages that take users to different locations depending on
 where they click.  To enable imagemaps first make sure the imagemap
 module is installed (its one of the default modules).  Next you need
 to uncomment the .map handler at the end of the srm.conf file.  Now
 all files ending in .map will be imagemap files.  Imagemap files map
 different areas on the image to separate links.  Apache uses map files
 in the standard NCSA format.  Here is an example of using a map file
 in a web page:


      <a href="/map/mapfile.map">
      <img src="picture.gif" ISMAP>
      </a>



 In this example mapfile.map is the mapfile, and picture.gif is the
 image to click on.

 There are many programs that can generate NCSA compatible map files or
 you can create them yourself.  For a more detailed discussion of
 imagemaps and map files see the Apacheweek feature at
 <http://www.apacheweek.com/features/imagemaps>.


 7.12.  SSI/XSSI

 Server Side Includes (SSI) adds dynamic content to otherwise static
 web pages.  The includes are embedded in the web page as comments.
 The web server then parses these includes and passes the results to
 the web server.  SSI can add headers and footers to documents, add
 date the document was last updated, execute a system command or a CGI
 script.  With the new eXtended Server Side Includes (XSSI) you can do
 a whole lot more.  XSSI adds variables and flow control statements
 (if, else, etc).  Its almost like having an programming language to
 work with.

 Parsing all HTML files for SSI commands would waste allot of system
 resources.  Therefore you need to distinguish normal HTML files from
 those that contain SSI commands.  This is usually done by changing the
 extension of the SSI enhanced HTML files.  Usually the .shtml
 extension is used.

 To enable SSI/XSSI first make sure that the includes module is
 installed.  Then edit srm.conf and uncomment the AddType and
 AddHandler directives for .shtml files.   Finally you must set Options
 Includes for all directories where you want to run SSI/XSSI files.
 This is done in the access.conf file.  Now all files with the
 extension .shtml will be parsed for SSI/XSSI commands.

 Another way of enabling includes is to use the XBitHack directive.  If
 you turn this on it looks to see if the file is executable by user.
 If it is and Options Includes is on for that directory, then it is
 treated as an SSI file.  This only works for files with the mime type
 text/html (.html .htm files).  This is not the preferred method.


 There is a security risk in allowing SSI to execute system commands
 and CGI scripts.  Therefore it is possible to lock that feature out
 with the Option IncludesNOEXEC instead of Option Includes in the
 access.conf file.  All the other SSI commands will still work.

 For more information see the Apache mod_includes documentation that
 comes with the source.  It is also available on the website at
 <http://www.apache.org/docs/mod/mod_include.html>.

 For a more detailed discussion of SSI/XSSI implementation see the
 Apacheweek feature at  <http://www.apacheweek.com/features/ssi>.

 For more information on SSI commands see the NCSA documentation at
 <http://hoohoo.ncsa.uiuc.edu/docs/tutorials/includes.html>.

 For more information on XSSI commands goto
 <ftp://pageplus.com/pub/hsf/xssi/xssi-1.1.html>.


 7.13.  Module system

 Apache can be extended to support almost anything with modules.  There
 are allot of modules already in existence.  Only the general interest
 modules are included with Apache.   For links to existing modules goto
 the

 Apache Module Registry at <http://www.zyzzyva.com/module_registry/>.

 For module programming information goto
 <http://www.zyzzyva.com/module_registry/reference/>



 8.  Web Server Add-ons

 Sorry this section has not been written yet.

 Coming soon: mSQL, PHP/FI, cgiwrap, Fast-cgi, MS frontpage extentions,
 and more.



 9.  Intranet Section



 In simple terms, the Intranet is the descriptive term being used for
 the implementation of Internet technologies within a corporate
 organisation, rather than for external connection to the global
 Internet. This implementation is performed in such a way as to
 transparently deliver the immense informational resources of an
 organisation to each individuals desktop with minimal cost, time and
 effort. This document attempts to explain in simple terms how to setup
 an Intranet using tools which are readily available and are generally
 costing little or are free.

 This document assumes that you already know how to install TCP/IP on
 your Linux server and connect it physically to your LAN using an
 Ethernet network card. This also assumes you have some basic knowledge
 of Netware, WinNT and Mac systems. The configuration of the Netware
 server has been shown using version 3.1x as the basis. You can also
 use INETCFG to achieve the same result. On the client side the
 discussion is with respect to Windows 3.1x, Windows for Workgroups and
 Win95, WinNT and the Apple PowerMac.

 I am using the private network addresses (RFC-1918) of 172.16.0.0 and
 172.17.0.0 only as examples. You may choose suitable addresses
 depending on your configuration.



             Linux       Netware      WFWG/WinNT
             Server      Server       Server
            172.16.0.1   172.16.0.2   172.16.0.3
               |           |           |    172.16.0.0
         ------+-----+-----+--------+--+--------------
                     |              | 172.16.0.254
                   W/S 1          Router
                 172.16.0.5         | 172.17.0.254
                                    |
                          ----------+-------+--------
                     172.17.0.0             |
                                          W/S 2
                                        172.17.0.5



 9.1.  What is required

 You will need the following software before attempting the
 installation.

 o  the HTTP server software which can be downloaded from OneStep NCSA
    HTTPd Downloader at
    http://hoohoo.ncsa.uiuc.edu/docs/setup/OneStep.html page.

 o  The Novell Netware Client available from HTTP://support.novell.com/
    (The TCP/IP files are included with the client).

 o  The Microsoft TCP/IP client available from
    HTTP://www.microsoft.com/

 o  The Apple MacTCP client available from HTTP://www.apple.com/

 o  WWW Browsers like Netscape at HTTP://home.netscape.com/ or MS
    Internet Explorer at HTTP://www.microsoft.com/ or NCSA Mosaic from
    http://www.ncsa.uiuc.edu/SDG/Software/Mosaic/NCSAMosaicHome.html


 9.2.  New versions of this document

 New versions of the Linux Intranet Server HOWTO will be periodically
 posted to comp.os.linux.announce and comp.os.linux.help. They will
 also be uploaded to various Linux FTP sites, including
 sunsite.unc.edu.

 The Latest version of this document is available in HTML format at
 http://www.linuxports.com/


 9.3.  Feedback

 If you have questions or comments about this document, please feel
 free to mail Mr. Poet, at [email protected]. Suggestions, criticism
 and mail are always welcome. If you find a mistake with this document,
 please let me know so I can correct it in the next version. Thanx.
 10.  Install the HTTP server

 When you download the server you have two options: To get the source
 and compile it yourself, or get the precompiled binaries. The
 precompiled binaries for Linux (ELF) version are available at NCSA but
 not the older versions.


 10.1.  Preparation before downloading

 The server at NCSA will guide you through the steps for configuration
 options and prepare the various files for you. But before you attempt
 to download HTTPd be prepared with answers to the following questions


 10.1.1.  The Operating System

 First, you must choose whether to download the source or a pre-
 compiled version of the software. If your particular system doesn't
 appear in the menu, then you will have to get the default source, and
 compile it yourself.

 To check the version of your Linux go to the command prompt on your
 Linux machine and type


        linux:~$  uname -a



 which will respond with a line which looks similar to this


         linux:~$  uname -a
         Linux linux 2.0.29 #4 Tue Sep 13 04:05:51 CDT 1994 i586
         linux:~$



 The version of Linux is 2.0.29.

 The remaining parameters can be specified before downloading or
 configured later by modifying the file srm.conf in the
 /usr/local/etc/httpd/conf directory. The names of the actual
 directives that appear in the file httpd.conf are shown in brackets.
 The only exception is the directive DocumentRoot which appears in the
 file srm.conf


 10.1.2.  Process type (ServerType)

 This specifies how your machine will run your HTTPd server. The
 preferred method is "standalone". This makes the HTTP daemon to be
 running constantly. If you choose to load HTTPd under "inetd", the
 server binary will be reloaded into memory for every request, which
 may slow your server down.


 10.1.3.  Binding Port (Port)

 This specifies which port of your machine that the HTTPd daemon will
 bind to and listen for HTTP requests. If you can login as "root", use
 the default setting of 80.  Otherwise choose a setting between 1025
 and 65535.
 10.1.4.  Server user identity (User)

 This is the user id the server will change to when answering requests
 and acting on files.This question needs to be answered only if you are
 running the server as "standalone".  If you are someone without root
 permissions, just use your own login name. If you are system
 administrator, you might want to create a special user so you can
 control file permissions.


 10.1.5.  Server group identity (Group)

 This is the group id the server will change to when answering requests
 and acting on files. This is similar to Server User identity and is
 applicable only if you are running the server as standalone.

 If you do not have root permissions, just use the name of your primary
 group. You can find out your group by typing groups at the Linux
 command prompt.


 10.1.6.  Server administrator email address (ServerAdmin)

 This is the email address that the user should send an email message
 to when reporting a problem with the server. You can put your personal
 e-mail address.


 10.1.7.  Location of server directory (ServerRoot)

 This is where the server resides on your system. If you have root
 permissions leave it in its recommended location /usr/local/etc/httpd.
 If you cannot login as root, choose a subdirectory in your home path.
 You can find out the path of your home directory with the pwd command.


 10.1.8.  Location of HTML files (DocumentRoot)

 This is where the HTML files to be served are located. The default
 location is /usr/local/etc/httpd/htdocs.  You could however set it to
 be the home directory of the special user you chose in Server user
 identity, or a subdirectory in your home directory if you can't login
 as root.

 When in doubt, use the default settings. Now that you have answers to
 the above questions you can Download NCSA HTTPd at
 http://hoohoo.ncsa.uiuc.edu/docs/setup/OneStep.html.  You should read
 the HTTPd Documentation at http://hoohoo.ncsa.uiuc.edu/docs/ before
 you attempt installation. If you are planning to compile the code then
 you need to modify the makefiles in each of the th ree directories
 support, src, cgi-src. If your version of Linux is already supported
 then you just have to type make linux at the top level directory (i.e.
 /usr/local/etc/httpd)


 10.2.  Compiling HTTPd

 Compiling is simple, just type make linux at the prompt in the server
 root directory.  Note: Users of pre-ELF Linux have to uncomment
 #define NO_PASS in file portability.h and set DBM_LIBS= -ldbm in the
 Makefile before compiling HTTPd.



 11.  Testing HTTPd

 After you have installed HTTPd, login as root and start it by typing
 httpd & . (assuming you have installed as standalone) You should now
 be able to see it in the list generated by ps. The simplest way to
 test HTTPd is by Telnet. At the Linux command prompt type


         linux:~$  telnet 172.16.0.1 80



 where 80 is the default port for HTTP. If you have configured "Port"
 as something different then type that number instead. You should get a
 response which looks like this


         Trying 172.16.0.1...
         Connected to linux.mydomain.
         Escape character is '^]'.



 Now if you type in any character and press Enter you should get a
 response similar to the one shown below.


         HTTP/1.0 400 Bad Request
         Date: Wed, 10 Jan 1996 10:24:37 GMT
         Server: NCSA/1.5
         Content-type: text/html

         <HEAD><TITLE>400 Bad Request < /TITLE> < /HEAD>
         <BODY><H1>400 Bad Request < /H1>
         Your client sent a query that this server could
         not understand.<P>
         Reason: Invalid or unsupported method.<P>
         < /BODY>



 Now we are ready to connect to this server using another PC and a WWW
 Browser.


 12.  Connecting to the Linux Server

 Please refer to the diagram shown in the chapter ``Introduction'' for
 the addressing scheme used. Workstation 1 (W/S1) is on network
 172.16.0.0 and can access the Linux server directly whereas
 Workstation 2 (W/S2) is on network 172.17.0.0 and needs to use the
 gateway (router) 172.17.0.254 to access the Linux box. This gateway
 information needs to be provided while configuring the clients only on
 W/S2. Netware refers to the gateway as 'ip_router'.

 I am using W/S2 to illustrate the client setup. To setup W/S1 just
 change the address 172.17.0.5 to 172.16.0.5 and ignore all references
 to the gateway/router.

 If you do not have a router you can skip the next section and proceed
 to


 o  ``Setup Netware Server'' if you use a Netware server.

 o  ``Setup MS Windows Client'' if you use the Microsoft Client.


 12.1.  Setup the Linux server

 You may skip this section if you do not have a router.

 You have to configure the Linux server to recognise the router thus
 allowing Workstation 2 to connect to the Web server. In order to setup
 the Linux server you should login as root. At the server prompt type


         route add gw default 172.16.0.254



 To use this gateway everytime you boot the Linux server edit the file
 /etc/rc.d/rc.inet1 and change the line containing the gateway defini-
 tion to GATEWAY = "172.16.0.254". Make sure the line for adding the
 gateway is not commented out.

 ALT: You can add routes to the networks on the other side of the
 router. This would be done as


         route add -net 172.17.0.0 gw 172.16.0.254



 To add this route everytime you boot Linux add the command to your
 /etc/rc.d/rc.local file.



 12.2.  Setup the Netware server

 In order to setup the Netware server you should have Supervisor
 permissions or atleast Console operator permissions. If these cannot
 be got, try asking your Network Administrator to help you with the
 setup.  At the Server enable the Ethernet_II frame type on the LAN by
 typing these commands or include them in the AUTOEXEC.ncf file.


         load NE2000 frame=Ethernet_II name=IPNET
         load TCPIP
         bind IP to IPNET addr=172.16.0.2 mask=FF.FF.FF.0



 You might have to specify the slot or board number while loading the
 NE2000 driver depending on your machine configuration. (eg: load
 NE2000 slot=3 frame=.....)


 12.3.  Setup the Netware Client

 On the PC you have the choice of Win3.1,WFWG or Win95. The
 installation procedure differs between Win95 and the older windows if
 you are using the 32bit client from Microsoft or Novell. If you are
 going to use the 16bit client, the procedure is the same and you can
 refer to the Windows 3.x installation instructions. For installing the
 32bit client for Win95 skip to ``Windows 95 installation''.


 12.3.1.  Windows 3.x

 If you are using Win3.1 or WFWG you can install the Netware Client
 (VLMs) and some additional files which are provided with the TCP/IP
 diskette, namely

 TCPIP.exe, VTCPIP.386, WINSOCK.dll and WLIBSOCK.dll

 Note that the WINSOCK.dll file is different from the ones provided
 with Win95 and Trumpet. Install the Netware Client with the support
 for windows. Copy VTCPIP.386, WINSOCK.dll and WLIBSOCK.dll to the
 SYSTEM directory and TCPIP.exe to the NWCLIENT directory. Now modify
 the STARTNET.bat in the NWCLIENT directory to


         lsl
         ne2000     ---> your network card driver
         c:\windows\odihlp.exe     ---->if you are using WFWG
         ipxodi
         tcpip        ---> add this line
         nwip         ---> if you use Netware/IP
         vlm



 Create a subdirectory (say) \NET\TCP and copy the files HOSTS,
 NETWORKS, PROTOCOLS and SERVICES from /etc on your Linux server or the
 directory SYS:ETC on your Netware server. Edit the copied HOSTS file
 to add the line for your new Linux server. This will enable you to
 refer to the Linux server as http://linux.mydomain/ instead of
 http://172.16.0.1/in your WWW browser


         127.0.0.1      localhost
         172.16.0.1      linux.mydomain



 Edit the NET.cfg file in NWCLIENT directory



    Link Driver NE2000
        port 300
        int 3
        MEM D0000
        FRAME Ethernet_802.2

 ; ---- add these lines ----

       FRAME Ethernet_II

    Protocol TCPIP
       PATH TCP_CFG C:\NET\TCP
       ip_address  172.17.0.5
       ip_netmask  255.255.255.0
       ip_router   172.17.0.254     ---> add the address of your gateway only
                                   ---> if you have to use this
                                   ---> gateway to reach your HTTP server

    Link Support
       MemPool 6192       ---> the minimum is 1024. Try with different values
       Buffers 10 1580    ---> this again can be fine tuned

 ;---------------------------------
 ; You may need to add lines like these if you are using Netware/IP
 ;
    NWIP
       NWIP_DOMAIN_NAME  mydomain
       NSQ_BROADCAST   ON
       NWIP1_1 COMPATIBILITY   OFF
       AUTORETRIES     1
       AUTORETRY SECS  10



 Edit the SYSTEM.ini file in the WINDOWS directory and add this entry
 for VTCPIP.386


         [386Enh]
         .....
         network=*vnetbios, vipx.386, vnetware.386, VTCPIP.386
         .....



 Reboot your PC, run STARTNET.bat and you can now use your favorite WWW
 browser to access your Web pages. You need not login to Netware and
 you don't have to run TCPMAN (if you use Trumpet Winsock).


 12.3.2.  Windows 95

 This section explains how to install the 32bit client on Win95.
 Firstly you must install the following


         Client for Netware Networks (from Microsoft or Novell)
         Microsoft TCP/IP Protocol
         Network Adapter



 To install these items, click on My Computer, Control Panel, Networks.
 Click Add. You will now be in a window that displays Client, Adapter,
 Protocol and Service.  To install the Client for Netware Networks:


        1. DoubleClick on Client
        2. Click on Microsoft or Novell
        3. DoubleClick on Client for Netware Networks



 To install the TCP/IP Protocol:


        1. DoubleClick on Protocol
        2. Click on Microsoft
        3. DoubleClick on TCP/IP



 Windows 95 by default installs several other protocols automatically.
 Remove them by clicking on them and clicking the Remove button.
 Typically Win95 installs the Microsoft NetBeui protocol, and IPX/SPX
 compatible protocol. You can delete the NetBEUI protocol, but you will
 need the IPX/SPX protocol if you wish to login to the Netware Server.

 To setup TCP/IP click on TCP/IP, click on Properties, click on the tab
 IP address


        Enter your IP address in the "Specify an IP address "
                              box as 172.17.0.5
        In the Subnet Mask box enter 255.255.255.0



 select the tab Gateway


        Enter your gateway (router) address in the box New gateway
           as 172.17.0.254
        Click the Add button



 The gateway address should now appear under the installed gateways
 box. Now Click OK.

 You should get a message to reboot. Do so. You should now be able to
 use the Browser to connect to your HTTP Server.


 12.4.  Setup Microsoft Client

 If you are using the Microsoft Client for accessing your network, then
 this section details how to install TCP/IP for

 o  ``Windows for Workgroups (WFWG)''

 o  ``Windows 95''


 o  ``Windows NT 4.0''

    Note: To enable you to refer to the Linux server as
    http://linux.mydomain/ instead of http://172.16.0.1/ in the WWW
    browser and all your intranet commands you need to edit the hosts
    file. You can add more entries for each of your other hosts
    (Netware, Unix, WinNT) as well. The Windows family keeps its HOSTS
    file in \WINDOWS or in \WINDOWS\SYSTEM depending on the version.
    Edit this file and add a line for your Linux server as:


         127.0.0.1      localhost
         172.16.0.1     linux.mydomain

         172.16.0.2     netware.mydomain
         172.16.0.3     winNT.mydomain
         172.16.0.5     ws_1



 12.4.1.  Windows for Workgroups

 This section explains how to install the 32bit client on WFWG. Firstly
 you must download the TCP/IP drivers for Windows from Microsoft. The
 current version is 3.11b and is available at ftp://ftp.microsoft.com
 or other sites as tcp32b.exe. Make sure that you have load Win32s
 before trying to load the TCP/IP-32bit driver.

 Having expanded the TCP/IP files into a temporary directory (say
 C:\TEMP), check your \WINDOWS\SYSTEM directory for copies of
 OEMSETUP.INF. If there are any, rename them. Now copy the OEMSETUP.INF
 file from the TEMP directory to the \WINDOWS\SYSTEM directory. If you
 have loaded any other TCP/IP stacks on your system, please remove them
 before you proceed.

 Start Network Setup or Windows Setup/Change Network settings


          Click the Networks button
          Click Install Microsoft Windows Network.
            Choose support for additional networks (if required)
          Click OK



 You should be prompted for your network adapter - select the appropri-
 ate one. If you are not prompted, then


           Click the Adapter button
               select an adapter (say NE2000)
               Click OK
           Click the Protocol button
               select the MS TCP/IP-32 protocol
               click OK



 You will now be prompted to configure the TCP/IP protocol stack. You
 can always reconfigure this by highlighting the TCP/IP protocol shown
 in the box Adapters and clicking the Setup button.

    In the IP address box enter 172.17.0.5
    In the Subnet Mask box enter 255.255.255.0
    Enter your gateway (router) address in the box default gateway
      as 172.17.0.254



 Click OK. The computer will ask you to restart. Do so. You should now
 be able to use the Browser to connect to your HTTP Server.


 12.4.2.  Windows 95

 This section explains how to install the 32bit client for Microsoft on
 Win95. Firstly you must install the following


         Client for Microsoft Networks
         Microsoft TCP/IP Protocol
         Network Adapter



 To install these items, click on My Computer, Control Panel, Networks.
 Click Add. You will now be in a window that displays Client, Adapter,
 Protocol and Service.  To install the Client for Microsoft Networks:


        1. DoubleClick on Client
        2. Click on Microsoft
        3. DoubleClick on Client for Microsoft Networks



 To install the TCP/IP Protocol:


        1. DoubleClick on Protocol
        2. Click on Microsoft
        3. DoubleClick on TCP/IP



 Windows 95 by default installs several protocols automatically. Remove
 them by clicking on them and clicking the Remove button. Typically
 Win95 installs the Microsoft NetBeui protocol.

 To setup TCP/IP click on TCP/IP, click on Properties, click on the tab
 IP address


        Enter your IP address in the "Specify an IP address "
                              box as 172.17.0.5
        In the Subnet Mask box enter 255.255.255.0



 select the tab Gateway



   Enter your gateway (router) address in the box New gateway
      as 172.17.0.254
   Click the Add button



 The gateway address should now appear under the installed gateways
 box. Now Click OK.

 You should get a message to reboot. Do so. You should now be able to
 use the Browser to connect to your HTTP Server.


 12.4.3.  Windows NT

 This section details how to Install the TCP/IP client for WinNT 4.0.
 Start Control Panel/ Network


          Select the Adapter tab.
              Click Add to add a new adapter (if you don't have one)



 You should be prompted for your network adapter - select the appropri-
 ate one. To add the protocols.


          Select the Protocols tab
              Click Add
              Select the TCP/IP protocol
              Click OK



 You will now be prompted to configure the TCP/IP protocol stack. You
 can always reconfigure this by highlighting the TCP/IP protocol and
 clicking the Properties button.


         Select the tab IP Address
            Mark the checkbox 'Specify an IP address'
            In the IP address box enter 172.17.0.5
            In the Subnet Mask box enter 255.255.255.0
            Enter your gateway (router) address in the box Default Gateway
                as 172.17.0.254



 Click OK. The computer will ask you to restart. You can now use any
 Browser to connect to your HTTP Server.


 12.5.  Setup TCP/IP on Macintosh

 If you are using the Macintosh for accessing your network, then this
 section details how to install MacTCP for the PowerMacs.

 Note: To enable you to refer to the Linux server as
 http://linux.mydomain/ instead of http://172.16.0.1/ in the WWW
 browser and all your intranet commands you need to edit the hosts
 file. The format of the hosts file is different from the one used in
 Unix. The Mac hosts file is based on RFC-1035. You can add more
 entries for each of your other hosts (Netware, Unix, WinNT) as well.
 The MacOS  keeps its HOSTS file in the Preferences folder under the
 System folder. Edit this file and add a line for your Linux server as:


        linux.mydomain    A  172.16.0.1

        netware.mydomain  A  172.16.0.2
        winNT.mydomain    A  172.16.0.3
        ws_1              A  172.16.0.5



 12.5.1.  MacTCP

 This section explains how to install MacTCP. Firstly you must download
 the MacTCP files from Apple or install it from the Internet Connection
 CD.  To configure MacTCP, click the Apple Menu/ Control Panels/
 TCP/IP. In the screen change the setting for 'Connect via:' to
 'Ethernet'

 Change the 'Configure' setting to 'Manually'


        In the IP address box enter 172.17.0.5
        In the Subnet Mask box enter 255.255.255.0
        Enter your gateway (router) address in the box
          Router address as 172.17.0.254



 Click OK. You should now be able to use the Browser to connect to your
 HTTP Server.



 13.  Setting up the Intranet

 An Intranet cannot be complete without sharing the resources on the
 different platforms. You will need support for other filesystems, so
 that you can access the data available on them. This document provides
 instructions to connect Linux to the following popular filesystems.

 o  ``NCP filesystem for Netware''

 o  ``SMB filesystem for Windows''

 o  ``NFS filesystem for Unix''

 These filesystems can be compiled into the Linux kernel or added as
 modules, depending on the version of Linux. If you are not familiar
 with compiling the kernel you can refer to the Kernel HOWTO
 http://sunsite.unc.edu/mdw/HOWTO/Kernel-HOWTO.html and the Module
 HOWTO http://sunsite.unc.edu/mdw/HOWTO/Module-HOWTO.html for compiling
 the kernel with modules.



 13.1.  NCPFS

 To share the files on the Netware server you will need support for NCP
 (ncpfs). NCPFS works with kernel version 1.2.x and 1.3.71 upwards. It
 does not work with any earlier 1.3.x kernel. It cannot access the NDS
 database in Netware 4.x, but can make use of the bindery. If you are
 using Netware 4.x you can enable bindery support for specific
 containers using the command Set Bindery Context at the console as:



        set Bindery Context = CORP.MYDOM;WEBUSER.MYDOM



 In the above example two containers have bindery support enabled.

 You will need to download the NCP filesystem utilities using the URL
 ftp://sunsite.unc.edu/pub/Linux/system/filesystems/ncpfs/ncpfs.tgz
 (currently ncpfs-2.0.10) from Sunsite.


 13.1.1.  Installation

 To install the ncpfs utilities, type


         zcat ncpfs.tgz | tar xvf -



 to expand the files into its own directory. In this case you will get
 a directory ncpfs-2.0.10 Change your directory to this ncpfs directory
 before proceeding with the installation. Read the README and edit the
 Makefile if necessary.

 The installation of ncpfs depends on the kernel version you are using.
 For kernel 1.2, you should simply type 'make'. Subsequently typing
 'make install' will install the executables and man pages.

 If you use Kernel 1.3.71 or later, you might have to recompile your
 kernel. With these kernels, the kernel part of ncpfs is already
 included in the main source tree. To check if the kernel needs to be
 recompiled type


         cat /proc/filesystems



 It should show you a line saying that the kernel knows ncpfs.

 If ncpfs is not there, you can either recompile the kernel or add
 ncpfs as a module. For recompiling the kernel you should type 'make
 config' and when it asks you for

 The IPX protocol (CONFIG_IPX) [N/y/?]

 simply answer 'y'. Probably you do not need the full internal net that
 you are asked for next. Once the kernel is successfully installed,
 reboot, check /proc/filesystems and if everything is OK proceed with
 the installation of the ncpfs utilities. Change directory to the
 location holding your downloaded ncpfs files, and type 'make'. After
 the compilation is finished type 'make install' to install the various
 utilities and man pages.


 13.1.2.  Mounting NCPFS

 To check the installation type


         ipx_configure --auto_interface=on --auto_primary=on

      ....wait for 10 seconds and type

         slist



 You should be able to see a list of your Netware servers. Now we are
 ready to share files from the Netware server.

 Suppose we need to access HTML files from directory \home\htmldocs on
 volume VOL1: on the server MYDOM_NW, I recommend that you create a new
 user (say) 'EXPORT' with password 'EXP123' on this server to whom you
 grant appropriate access rights to this directory using SYSCON or
 NWADMIN.

 On the Linux machine create a new directory /mnt/MYDOM_NW. Now type
 the command


         ncpmount -S MYDOM_NW -U EXPORT -P EXP123 /mnt/MYDOM_NW



 to mount the netware file system. Typing the command


         ls /mnt/MYDOM_NW/vol1/home/htmldocs



 will show you a list of all the files in MYDOM_NW/VOL1:\HOME\HTMLDOCS
 (using Netware file notation). If you have any problems please read
 the IPX HOWTO at http://sunsite.unc.edu/mdw/HOWTO/IPX-HOWTO.html for
 more insights into the IPX system.



 13.2.  SMBFS

 To share the files on the Windows server you will need support for SMB
 (smbfs).

 You will need to download the SMB filesystem utilities from
 ftp://sunsite.unc.edu/pub/Linux/system/filesystems/smbfs/smbfs.tgz
 (currently smbfs-2.0.1) from Sunsite.


 13.2.1.  Installation

 To install the smbfs utilities, type


         zcat smbfs.tgz | tar xvf -



 to expand the files into its own directory. In this case you will get
 a directory smbfs-2.0.1 Change your directory to this smbfs directory
 before proceeding with the installation. Read the README and edit the
 Makefile if necessary.

 The installation of smbfs depends on the kernel version you are using.
 For kernel 1.2, you should simply type 'make'. Subsequently typing
 'make install' will install the executables and man pages.

 If you use Kernel 2.0 or later, you might have to recompile your
 kernel. With these kernels, the kernel part of smbfs is already
 included in the main source tree. To check if the kernel needs to be
 recompiled type


         cat /proc/filesystems



 It should show you a line saying that the kernel knows smbfs.

 If smbfs is not there, you can either recompile the kernel or add
 smbfs as a module. For recompiling the kernel you should type 'make
 config' and when it asks you for adding SMB filesystem support simply
 answer yes. Once the kernel is successfully installed, reboot, check
 /proc/filesystems and if everything is OK proceed with the
 installation of the smbfs utilities. Change directory to the location
 holding your downloaded smbfs files, and type 'make'. After the
 compilation is finished type 'make install' to install the various
 utilities and man pages.


 13.2.2.  Mounting SMBFS

 In our example let us assume that the WinNT server is called
 'MYDOM_NT' and is sharing its directory C:\PUB\HTMLDOCS with a share
 name of 'HTMLDOCS' without a password. On the Linux machine create a
 new directory /mnt/MYDOM_NT. Now type the command


         smbmount //MYDOM_NT/HTMLDOCS /mnt/MYDOM_NT -n



 to mount the SMB (windows share) file system. If this does not work
 try


         smbmount //MYDOM_NT/COMMON /mnt/MYDOM_NT -n -I 172.16.0.3



 Typing the command


         ls /mnt/MYDOM_NT



 will show you a list of all the files in bsol;bsol;MYDOM_NT\PUB\HTML-
 DOCS (using Windows file notation).

 13.3.  NFS

 First you will need a kernel with the NFS file system either compiled
 in or available as a module.

 Suppose you have a Unix host running NFS with the name MYDOM_UNIX and
 an IP address of 172.16.0.4. You can check the directories that are
 being exported (shared) by this host by typing the command



         showmount -e 172.16.0.4



 Once we know the exported directories you can mount them by entering a
 appropriate mount command. I recommend that you create a subdirectory
 under '/mnt' (say) 'MYDOM_UNIX' and use that as your mount point.



         mount -o rsize=1024,wsize=1024 172.16.0.4:/pub/htmldocs /mnt/MYDOM_UNIX



 The rsize and wsize may have to be changed depending on your environ-
 ment.

 If you have any problems please read the NFS HOWTO at
 http://sunsite.unc.edu/mdw/HOWTO/NFS-HOWTO.html for more insights into
 the NFS system.



 14.  Accessing the Web

 Now that we have setup the HTTP server, the clients and interconnected
 the Linux server with the other servers, we need to make some small
 adjustments on the Linux server to be able to access these mounted
 filesystems from the Web Browser.


 14.1.  Accessing the mounted filesystems

 To access the mounted directories in your HTML pages you have two
 methods:

 o  Create a link in DocumentRoot (/usr/local/etc/httpd/htdocs) to
    refer to the mounted directory as


         ln -s /mnt/MYDOM_NW/vol1/home/htmldocs netware
                           or

         ln -s /mnt/MYDOM_NT    winNT
                           or

         ln -s /mnt/MYDOM_UNIX    unix



 o  to edit the file srm.conf in your /usr/local/etc/httpd/conf
    directory and add a new alias.
        # Alias fakename realname
        Alias /icons/     /usr/local/etc/httpd/icons/

        # alias for netware server
        Alias /netware/   /mnt/MYDOM_NW/vol1/home/htmldocs/
        Alias /winNT/     /mnt/MYDOM_NT/
        Alias /unix/      /mnt/MYDOM_UNIX



 And restart your HTTPd. You can access the documents on the netware
 server by referring to them as http://linux.mydomain/netware/index.htm
 for the netware files and similar notations for the others.


 14.2.  Connecting to the Internet

 You can finally connect your Intranet to the Internet to access E-Mail
 and all the wonderful information out there. I propose to write a
 brief note on how to do this in a future revision. Detailed
 explanations are available in the ISP Hookup HOWTO from
 http://sunsite.unc.edu/mdw/HOWTO/ISP-Hookup-HOWTO.html and Diald mini
 HOWTO at http://sunsite.unc.edu/mdw/HOWTO/mini/Diald for setting up
 these connections.


 14.3.  Other uses

 The HTTP server can be used in the office to provide transparent
 access to information residing on different servers, at several
 locations and directories. The data can be simple documents in Word,
 Lotus spreadsheets, or complex databases.

 The application of this technology is being typically used as
 follows:-


 o  Publishing corporate documents

    These documents can include newsletters, annual reports, maps,
    company facilities, price lists, product information literature,
    and any document which is of value within the corporate entity.

 o  Access into searchable directories

    Rapid access to corporate phone books and the like. This data can
    be mirrored at a Web site or, via CGI scripts, the Web server can
    serve  as a gateway to back-end pre-existing or new applications.
    This means that, using the same standard access mechanisms,
    information can be made more widely available and in a simpler
    manner.  This means that it can be used to create an interface with
    RDBMS  like ORACLE and SYBASE for generating real-time information.
    Here is a list of links to such sites on the Web.

    - Web Access -
    http://cscsun1.larc.nasa.gov/~beowulf/db/web_access.html - CGI
    gateways - HTTP://www.w3.org/hypertext/WWW/RDBGate/Overview.html


 o  Corporate/Department/Individual pages

    As cultures change within organistions to the point where even each
    department moves towards their own individual mission statements,
    the Intranet technology provides the ideal medium to communicate
    current information to the Department or Individual. Powerful
    search engines provide the means for people to find the group or
    individual who has the answers to the continuous questions which
    arise in the normal day-to-day course of doing business.

 o  Simple Groupware applications

    With HTML forms support, sites can provide sign-up sheets, surveys
    and simple scheduling.

 o  Software distribution

    Administrators can use the Intranet to deliver software and up-
    dates on-demand to users across the corporate network . This can be
    done with 'Java' which allows the creation and transparent
    distribution of objects on-demand rather than just data or
    applications. This is indeed possible more easily with the newer
    versions of Linux which has builtin support for Java.

 o  Mail

    With the move to the use of Intranet mail products with standard
    and simple methods for attachment of documents, sound, vision and
    other multimedia between individuals, mail is being pushed further
    forward as a simple, de facto communications method. Mail is
    essentially individual to individual, or individual to small group,
    communication. Several utilities are available on the Linux
    platform to setup an E-mail system like sendmail, pop3d, imapd.


 o  User Interface

    The Intranet technology is evolving so rapidly that the tools
    available, in particular HTML, can be used to dramatically change
    the way we interface with systems. With HTML you can build an
    Interface which is only limited by the creators imagination. The
    beauty about using Intranet technologies for this is that it is so
    simple. Clicking a hyperlink from HTML can take you to another
    page, it could ring an alarm, run a yearend procedure or anything
    else that a computer program can do.


 15.  More things to do

 Here is a list of other interesting things to do with your Linux
 Intranet server. All the software mentioned below is freeware or
 shareware.

 o  Browse the Linux server using Network Neighbourhood in Win95/ NT;
    Setup a WINS like NBT server. Check out the SAMBA Web page at
    http://lake.canberra.edu.au/pub/samba/samba.html

 o  Implement a search engine on your Intranet. Connect to ht://Dig at
    http://htdig.sdsu.edu/

 o  Use CUSeeMe by setting up a local reflector. Refer to their page at
    Cornell http://cu-seeme.cornell.edu/

 o  Setup Web Conferencing. Use COW from http://thecity.sfsu.edu/COW/

 o  Deploy a SQL database. Refer to the mSQL Home page at
    http://Hughes.com.au/

 o  Setup FTP,Gopher,Finger,Bootp servers on the Netware server. Get
    them at http://mft.ucs.ed.ac.uk/


 o  Emulate a Netware server. Check out the NCP Utilities at
    ftp://sunsite.unc.edu/pub/Linux/system/filesystems/ncpfs/

    If you find other interesting things to do with your Linux Intranet
    server, please feel free to mail me.


 16.  Credits and Legalities



 16.1.  Thanks

 Thanks to the people at NCSA for providing such excellent
 documentation, David Anderson and all others for trying out this HOWTO
 and sending in their comments. The details on Netware/IP are courtesy
 Romel Flores ([email protected]).



 16.2.  Copyright information

 This document is copyrighted (C) 1999 {Poet} and LinuxPorts, it is
 distributed under the following terms:


 o  Linux HOWTO documents may be reproduced and distributed in whole or
    in part, in any medium physical or electronic, as long as this
    copyright notice is retained on all copies. Commercial
    redistribution is allowed and encouraged; however, the author would
    like to be notified of any such distributions.

 o  All translations, derivative works, or aggregate works
    incorporating any Linux HOWTO documents must be covered under this
    copyright notice.  That is, you may not produce a derivative work
    from a HOWTO and impose additional restrictions on its
    distribution. Exceptions to these rules may be granted under
    certain conditions; please contact the Linux HOWTO coordinator at
    the address given below.

 o  If you have questions, please contact Tim Bynum, the Linux HOWTO
    coordinator, at [email protected]  Finger for phone
    number and snail mail address.



 17.  For further reading



 17.1.  O'Reilly & Associates Books

 In my humble opinion O'Reilly & Associates make the best technical
 books on the planet.  They focus mainly on Internet, Unix and
 programming related topics. They start off slow with plenty of
 examples and when you finish the book your an expert.  I think you
 could get by if you only read half of the book.  They also add some
 humor to otherwise boring subjects.

 They have great books on HTML, PERL, CGI Programming, Java,
 JavaScript, C/C++, Sendmail, Linux and much much more.  And the fast
 moving topics (like HTML) are updated and revised about every 6 months
 or so.  So visit the O'Reilly & Associates <http://www.ora.com/> web
 site or stop by your local book store for more info.


 And remember if it doesn't say O'Reilly & Associates on the cover,
 someone else probably wrote it.


 17.2.  Internet Request For Comments (RFC)


 o  RFC1866 written by T. Berners-Lee and D. Connolly, "Hypertext
    Markup Language - 2.0", 11/03/1995

 o  RFC1867 writtenm by E. Nebel and L. Masinter, "Form-based File
    Upload in HTML", 11/07/1995

 o  RFC1942 written by D. Raggett, "HTML Tables", 05/15/1996

 o  RFC1945 by T. Berners-Lee, R. Fielding, H. Nielsen, "Hypertext
    Transfer Protocol -- HTTP/1.0", 05/17/1996.

 o  RFC1630 by T. Berners-Lee, "Universal Resource Identifiers in WWW:
    A Unifying Syntax for the Expression of Names and Addresses of
    Objects on the Network as used in the World-Wide Web", 06/09/1994

 o  RFC1959 by T. Howes, M. Smith, "An LDAP URL Format", 06/19/1996

You are viewing proxied material from ftp.icm.edu.pl. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.