Getting Up and Running with Apache 2.0 on Linux

LinuxWorld NYC 2004: New York City, January 20, 2004.

Getting Up and Running with Apache 2.0 on Linux Introduction (Slide Two)

i am who i am who i am who am i requesting some enlightenment could I have been anyone other than me?"

-from Dancing Nancies, Dave Matthews Band

Introduction (Slide Three)

  • the topic at hand: apache 2.0
    • what we'll cover in the next 3 hours
      • history of
      • make up of
      • getting our hands on
      • what's on hand
      • configuring, building
      • configuring, migrating to
      • adding third-party modules
      • fine tuning
      • future of

Introduction (Slide Four)

  • the topic at hand: apache 2.0
    • what we won't cover
    • detailed changes in apache source code
    • detailed changes in apache APIs
    • detailed workings of apache modules
    • security precautions
    • apache in relation to a specific linux distribution

Introduction (Slide Five)

  • the topic at hand: apache 2.0
    • how this hands-on lab will work
      • build, configure latest apache 2.0 release on our linux laptops 2, 15 minute breaks
      • workbook, which contains the a copy of slides you see and additional information such as notes, comments and resources that further illuminate the topic at hand.
      • with 3 hours to spare, feel free to stop and ask questions.

Apache 2.0: A History (Slide Six)

"history is the version of past events that people have decided to agree upon"

-Napoleon Bonaparte

Apache 2.0: A History (Slide Seven)

  • plans for apache 2.0 date as far back as the summer of 1996

Apache 2.0: A History (Slide Eight)

  • in june 1997 an official set of requirements for rewriting the core apache code was put forth some of the issues discussed at the time included:
    • portability
    • scalability
    • configuration
    • I/O filtering

Apache 2.0: A History (Slide Nine)

  • in june 1998 the apache core developers met for the first time to discuss the organizational structure of the apache group as well

Apache 2.0: A History (Slide Ten)

  • in january 2000, 2.0 became the primary development branch for the apache web server

Apache 2.0: A History (Slide Eleven)

    the first apache 2.0 alpha was launched at the apachecon 2000 conference in march 2000

Apache 2.0: A History (Slide Twelve)

  • the first apache 2.0 beta was launched at the apacheCon 2001 conference in march 2001

Apache 2.0: A History (Slide Thirteen)

  • apache 2.0 "went gold" with the first general availability release of version 2.0.35 in april 2002

Apache 2.0: A History (Slide Fourteen)

  • in november 2002, it was decided that development of new features would take place in a 2.1 development tree

Apache 2.0: A History (Slide Fifteen)

  • as of 17th of nov 2004 the latest stable release of apache 2.0 apache 2.0.48 on 29th of oct 2003

New and Improved (Slide Sixteen)

"but wait, there's more!"

-from any given infomercial

New and Improved (Slide Seventeen)

  • new process support with multiple-processing modules (MPMs)
    • options vary depending on platform
    • on linux
      • prefork
      • prefork with threads (worker)
      • perchild

New and Improved (Slide Eighteen)

  • the Apache Portable Runtime (APR) was introduced as the abstraction layer that masks the call differences of various systems
  • the modules' APIs has introduced several enhancements to provide more flexibility in handling callback routines and data structures

New and Improved (Slide Nineteen)

  • I/O filtering
    • one module can modify the results of another
    • filters can be configured within the config file
    • modules can add filters based on request
  • the configuration file has changed
    • read once and stored in memory
    • easy to convert to XML
    • divided into multiple files

New and Improved (Slide Twenty)

  • apache 2.0 also supports the new internet protocol, IPv6
  • the apache group's distribution of apache 2.0 now includes additional modules such as mod_ssl, mod_dav, mod_cgid

Getting Apache 2.0 (Slide Twenty One)

"it's rather like going to buy coke, complaining to the coca-cola company that the drink was too sour, then finding that your supermarket was adding lemon"

-"Vendor patches to Apache" Apache Week, 3rd March 2003

Getting Apache 2.0 (Slide Twenty Two)

  • from linux distribution
    • Gentoo
    • Debian
    • Red Hat
    • S.u.S.E.

Getting Apache 2.0 (Slide Twenty Three)

Getting Apache 2.0 (Slide Twenty Four)

"the last thing anyone in the apache community seems to want is to compromise apache's open-source roots and cause fragmentation of a web server that is extremely popular. at the same time, most distributors want to make money on the software and fine-tune it to run with their linux distributions. walking that fine line is no easy task"

-"Patching Apache" NewsFactor Network, 21st May 2003

Building Apache 2.0 (Slide Twenty Five)

"truly, you have a dizzying intellect"

"wait 'til I get going"

- exchange from The Princess Bride

Building Apache 2.0 (Slide Twenty Six)

  • getting our bearings
    • apache file layout
      • no independent src/ directory
      • docs, config files, sample cgi scripts and error documents reside in docs/ directory
      • modules, os related items modules/ and os/ directories

Building Apache 2.0 (Slide Twenty Seven)

  • module-go-round:
    • gone
    • mod_auth_db
    • mod_digest
    • mod_log_agent
    • mod_log_referer
  • new
    • mod_cache
    • mod_auth_cgid
    • mod_charset_lite
    • mod_dav
    • mod_deflate
    • mod_ext_filter
    • mod_file_cache
    • mod_isapi
    • mod_ssl
    • mod_suexec

Building Apache 2.0 (Slide Twenty Eight)

  •  
    • server path is: /usr/local/apache2
    • modules as listed with ./httpd -l
      • core.c
      • mod_access.c
      • mod_auth.c
      • mod_include.c
      • mod_log_config.c
      • mod_env.c
      • mod_setenvif.c
      • prefork.c
      • http_core.c
      • mod_mime.c
      • mod_status.c
      • mod_autoindex.c
      • mod_asis.c
      • mod_cgi.c
      • mod_negotiation.c
      • mod_dir.c
      • mod_imap.c
      • mod_actions.c
      • mod_userdir.c
      • mod_alias.c
  • ./configure with no options:

Building Apache 2.0 (Slide Twenty Nine)

  • some additional options for --configure
    --enable-auth-digest
    --enable-echo
    --enable-mem-cache
    --enable-case-filter
    --enable-auth-ldap
    --enable-usertrack
    --enable-unique-id
    --enable-proxy
    --enable-info
    --enable-vhost-alias
    --enable-speling
    --enable-rewrite

Building Apache 2.0 (Slide Thirty)

  • dynamic or static modules
    DSO capability
    --enable-so

Building Apache 2.0 (Slide Thirty One)

./configure --prefix=/usr/local/apache-2.0.48 --enable-so
make
sudo make install

Internission (Slide Thirty Two)

15 minutes while everyone's system run build processes

Configuring Apache 2.0 (Slide Thirty Three)

"i am extraordinarily patient, provided I get my own way in the end"

- Margaret Thatcher

Configuring Apache 2.0 (Slide Thirty Four)

  • getting our bearings
    • apache file layout
        bin/
        cgi-bin/
        error/
        icons/
        lib/
        man/
        modules/
        build/
        conf/
        htdocs/
        include/
        logs/
        manual/

Configuring Apache 2.0 (Slide Thirty Five)

  • global environment configuration
    • this section has seen a large number of changes compared with apache 1.3

Configuring Apache 2.0 (Slide Thirty Six)

  • global environment configuration
    • The prefork MPM accepts the same directives as apache 1.3
      • StartServers
      • MinSpareServers
      • MaxSpareServers
      • MaxClients
      • MaxRequestsPerChild

Configuring Apache 2.0 (Slide Thirty Seven)

  • global environment configuration
    • the BindAddress and Port directives no longer exist
      • their functionality is now provided by a more flexible Listen and ServerName directives.

Configuring Apache 2.0 (Slide Thirty Eight)

  • global environment configuration
    • the AddModule and ClearModuleList directives no longer exist
    • the order of the LoadModule lines is no longer relevant

Configuring Apache 2.0 (Slide Thirty Nine)

  • global environment configuration
    • gone missing
      • ServerType
      • AccessConfig
      • ResourceConfig

Configuring Apache 2.0 (Slide Forty)

  • main server configuration
    • the section has seen little change between apache 1.3 and 2.0
    • this is not by chance, but design

Configuring Apache 2.0 (Slide Forty One)

  • main server configuration
    • logging
      • the following logging directives have been removed
        • AgentLog
        • RefererLog
        • RefererIgnore
      • agent and referrer logs are still available using the CustomLog and LogFormat directives

Configuring Apache 2.0 (Slide Forty Two)

  • main server configuration
    • directory indexing
      • FancyIndexing directive has now been removed
      • the new VersionSort option to the IndexOptions directive causes files containing version numbers to be sorted in a more natural way
      • the defaults for the ReadmeName and HeaderName directives have changed from README and HEADER to README.html and HEADER.html

Configuring Apache 2.0 (Slide Forty Three)

  • main server configuration
    • content negotiation
      • the CacheNegotiatedDocs directive now takes the argument on or off

Configuring Apache 2.0 (Slide Forty Four)

  • main server configuration
    • error documents
      • to use a hard-coded message with the ErrorDocument directive, the message should be enclosed in a pair of double quotation marks
      • greater flexibility with error documents

Configuring Apache 2.0 (Slide Forty Five)

  • virtual hosts configuration
    • the contents of all containers should be migrated in the same way as the main server section

Configuring Apache 2.0 (Slide Forty Six)

Modules Galore 2.0 (Slide Forty Seven)

"any sufficiently advanced technology is indistinguishable from magic

-Arthur C. Clarke

Modules Galore 2.0 (Slide Forty Eight)

  • hooks and optional functions
    • a hook represents an event that occurs in the course a connection
    • modules can register to participate in various hooks, indicating that they wish to have callback functions run during those stages of request processing

Modules Galore 2.0 (Slide Forty Nine)

  • hooks and optional functions
    • optional functions allow a module to register a function for use by another module with the core
    • if another module is interested in one of the registered functions, it checks with the core, retrieving a pointer to it if so

Modules Galore 2.0 (Slide Fifty)

  • hooks and optional functions
    • the benefits of this system are nicely illustrated in mod_include, which handles all of its SSI tags through an optional function

Modules Galore 2.0 (Slide Fifty One)

  • filters
    • chain actions on input or output, not connection events

Modules Galore 2.0 (Slide Fifty Two)

  • filters
    • example: Common Gateway Interface (CGI) scripts can generate server-parsed HTML documents which can then be processed by mod_include
    • the way this works is that each request is served by exactly one handler module followed by zero or more filter modules

Modules Galore 2.0 (Slide Fifty Three)

  • beginning with apache 2.0.42 the API will be kept stable
  • third party modules, php example
      ./configure ---with-apxs2=/usr/local/apache-2.0.48/bin/apxs --prefix=/usr/local/apache-2.0.48/php-4.3.2 --disable-cgi

Intermission (Slide Fifty Four)

15 minutes while everyone's system run build processes

Modules Galore 2.0 (Slide Fifty Five)

  • adding in PHP
      LoadModule php4_module modules/libphp4.so
      DirectoryIndex index.html index.php
      AddType application/x-httpd-php php

Fine Tuning Apache 2.0 (Slide Fifty Six)

"perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away"

-Antoine de Sain

Fine Tuning Apache 2.0 (Slide Fifty Seven)

  • MPM
    • best linux option: prefork
      • too early
        • third party modules, the source of apache's success, may not be threads aware
        • processing model is understood
      • other MPMs options in future
      • threading adds a bit of instability, even apache hybrid solution
      • threading probably won't provide large benefit for linux systems anyway

Fine Tuning Apache 2.0 (Slide Fifty Eight)

  • apache
    • keep an eye on invoking resource intensive configurations
      • HostNameLookup
      • FollowSysLinks
      • KeepAlive, KeepAliveTimeout, MaxKeepAlive
      • Log level option
      • session tracking
      • .htacess files, large httpd.conf files
      • modules such as
        • mod_status
        • mod_rewrite

Fine Tuning Apache 2.0 (Slide Fifty Nine)

  • linux
    • kernel
      • handle number processes/threads
    • filesystem
      • optimize block size
      • enable journaling

Fine Tuning Apache 2.0 (Slide Sixty)

  • apache
    • use apache bench for testing
      •  
        • 5,000 request, 500 at a time, using KeepAlive requests to our local server
      • ./ab -n5000 -c 500 -k localhost

The Road Ahead (Slide Sixty One)

for a successful technology, reality must take precedence over public relations, for nature cannot be fooled"

-Richard Feynman

The Road Ahead (Slide Sixty Two)

usage of apache 2.0.x

The Road Ahead (Slide Sixty Three)

usage of apache 1.3.x

The Road Ahead (Slide Sixty Four)

  • possible road blocks to adoption:
    • lack of support
      • hosting companies, modules
      • "if it ain't broke don't fix it"

The Road Ahead (Slide Sixty Five)

history of apache releases

The Road Ahead (Slide Sixty Six)

  • what's next for apache?
    • 2.1 developments
      • redesign of the MPMs code
      • redesign of AAA (authentication/authorization/access control) code

Review (Slide Sixty Seven)

  • what was covered
    • what is apache 2.0
    • building, configuring
    • running, troubleshooting

This Presentations (Slide Sixty Eight)