Giter Club home page Giter Club logo

debian-cd's Introduction

debian-cd
=========

Copyright 1999-2001 Raphaël Hertzog <[email protected]> and others,
2004-2010 Steve McIntyre <[email protected]>
This set of tools is licensed under the General Public License version
2 or any later version. You can find it in
/usr/share/common-licenses/GPL on a Debian GNU system.

Some of the ideas here (a loooong time ago) came from Steve McIntyre's
slink_cd script.

Thanks also to all the contributors on the debian-cd mailing list.

What is needed?
===============
Software:
- the apt-get (>= 0.3.11.1) tool
- perl (>= 5.004)
- bash (or another POSIX shell)
- make
- cpp
- mkisofs/genisoimage
- the perl Digest::MD5 module
- the perl Compress::Zlib module
- lynx (for text version of README.html) and todos from sysutils
  to convert docs to DOS format (although you can rip that out, too)
- if you want to generate jigdo files: jigdo-file (see below)
Other:
- apt-utils (for apt-ftparchive)
- lots of free space on your disks
- a Debian mirror, on a partition where you can write.
  If you can't write on it then you may try to use a symlink farm,
  but it's not the recommended way to build Debian CDs.


		-------------------------------
		- GENERATING DEBIAN CD IMAGES -
		-------------------------------


For the people that don't have time, here's the quick explanation
=================================================================

Edit the CONF.sh and change the PATHs for the mirror and so on.
$ sensible-editor CONF.sh
$ . CONF.sh
$ make distclean
$ make status
$ make official_images

However, you really should consider reading further for more information.
You can also take a look at build.sh and build_all.sh for an automated
way of building CD images. 

The script easy-build.sh offers the easiest way to build a specific image
or set of images, but is still very flexible and powerful. It is the
recommended tool for building test images and for people new to debian-cd.
See the file README.easy-build for further info.

How to build a CD set - step by step
====================================

If you haven't already, change to the /usr/share/debian-cd/ directory
(or, alternatively, set the variable BASEDIR in CONF.sh to point
there).

The process of building a CD is composed of the following steps:

- first configure what is needed in CONF.sh and source it in your shell:

    $ . CONF.sh

  The exported environment variables will be used by all the
  tools involved here (Makefiles, perl scripts, shell scripts).

  If you want to build CD images for more than one arch, you will
  have to build them one after the other (you may use a shell
  script for this).

  Note that the temporary dir must be on the same device as the
  mirror because debian-cd uses hardlinks for generating an image 
  tree. If you can't do this, you'll have to use a symlink farm.
  The symlink farm is explained at the end of this README.

  Keep in mind that the environment variables will stay in the
  environment after your debian-cd run, and may interfere with
  other program using the same variables (e.g. kernel-package).
  So if you want to be 100% safe, run debian-cd in a
  separate shell that you can leave after you're done.

- then we clean everything that may still be there from previous runs:

    $ make distclean

- then we initialize the temporary directory used for the build:

    $ make status

  If this has failed then this will be automatically launched:

    $ make correctstatus

  Note however that "make status" should never fail if it is
  used for building a CD set for the stable release...

- now you can decide what you want on your CDs

  Note that task files are always taken from the subdirectory in ./tasks/
  that matches the CODENAME environment variable. At the beginning of a
  build these "static" task files are copied to the working directory.
  During the build some additional task files - that are referenced from
  the static task files - are generated automatically using scripts from
  the ./tools directory.

  Examples:

    $ make packagelists TASK=Debian-generic COMPLETE=1

  or

    $ make packagelists TASK=Debian-kde COMPLETE=0

  or

    $ export NONFREE=1; make packagelists TASK=your-task-here COMPLETE=1

  or for something like an official image for the USA (without non-free):

    $ make packagelists COMPLETE=1

  .... take a look at the file tasks/* to see the options you can have :)

  You can change the behaviour of this command with the following
  variables:
  - if NONFREE is set, then packages from non-free will be allowed
    (NONFREE must be exported to all sub-shells)
  - if EXTRANONFREE is set, then non-free packages will be included
    on an extra CD (the last CD in fact). Don't use NONFREE and
    EXTRANONFREE at the same time!
    (EXTRANONFREE must be exported to all sub-shells)
  - if COMPLETE is set, all packages that are not listed in the
    selected task file will be included at the end
  - setting INSTALLER_CD will use an appropriate task file for
    building small CDs (businesscard and netinst)

- now, we'll start making temporary trees:

    $ make image-trees

  This will first work sort the list of packages for each architecture
  into order so that standard, required, important and base packages
  are placed first, then other packages will be added in the order
  given modulo dependency ordering. Once the sorted list is created,
  the different architecture lists will be merged (if more than one
  architecture is selected).

  Then the code will start laying out temporary directory trees for
  the CDs. In order, this includes the following steps:

  - Creating an empty directory layout
  - Generating an image label and volume ID and other metadata such as
    debian-installer information files
  - Add documentation and installation/upgrade tools
  - Add Release files and other archive metadata
  - Make the image bootable for the selected architecture(s)
  - Start generating the md5sum.txt file

  If you want to use boot-floppies built by yourself you can add
  a parameter BOOTDISKS=<dir> which specifies the directory where
  they are. Note that $BOOTDISKS/current must be a symlink to the
  real directory and it must follow the same setup as the FTP
  mirror. Your boot-floppies must also be on the same partition as
  your mirror and temporary dir (hardlinks are used here too).

  Once the disc tree has all of this start data, we start filling the
  directory trees with packages from the sorted list. The size of the
  image to be created is set using DISKTYPE in CONF.sh; if the
  standard sizes do not match what you're after, use DISKTYPE=CUSTOM
  and specify your own size using CUSTOMSIZE. The algorithm is simple
  for adding packages:

  - link the package into the temporary disc tree
  - append the metadata to the appropriate Packages or Sources file
  - add md5sum information for the added file(s) to the md5sum.txt file

  This continues until the temporary tree grows one package *too
  large* for the selected image size. At that point, we roll back the
  last set of changes associated with that package. Then:

  - check if the disc contains all the packages needed to install a base system
  - finish off the Release file, using the checksums of the
    Packages/Sources files we generated
  - finish off the md5sum.txt file

  Next, we continue to the next disc tree, using the same process:
  start it, copy packages in until they overflow, roll back and
  finish. And repeat. Each time a package is found to be too large to
  fit inside an image, it will be kept back and will (obviously) be
  the first package placed into the next disc tree.

- now we can create the images:

    $ make images

  If you don't have enough space for all images, you can generate 
  only one image (of the second CD for example) with:

    $ make image CD=2

  Note: here we use "make images", but you could as well use
  "make official_images" since the latter is the same as the former
  with some dependencies on targets that you already launched
  (make bootable packages sources).

- if you want to generate a MD5SUMS file with the md5sums of the
  images you can do it with:

    $ make imagesums

Official images
===============

If you use make official_images you're building CD images that have
the same properties than official CD images but they still doesn't
have the name of "Official Images". The name of the images is given
by setting the OFFICIAL and DEBVERSION environment variable (check
CONF.sh). 

Please never ever use the "Official" name for a CD image that you
built yourself. The only images that can be called "Official" are the
ones built by Debian itself and which are provided on Debian's
servers.

The default configuration shipped with this package will automatically
name the images "Unofficial". CD will work exactly in the same way
with all Debian tools, only the label is different. That means you
can use build.sh and build_all.sh to build your "Unofficial" images
without modifying anything.


Local packages
==============

If you provide some custom made packages and you want to put them on
Debian CD set you can do it. Simply put your packages in
$MIRROR/dists/$CODENAME/local/binary-$ARCH/<section>.
The organization of this sub-tree is the same than what you can find
in the main, contrib or non-free part. You may use different section
names if you want. Be sure to create Packages files (and Sources.gz if you
include sources).

You can also put your local packages under $MIRROR/pool/local (just a new
facility for people who don't want packages under dists/). To include local
packages, the LOCAL environment variable must be set to "1" while building
the CDs.

You can also set the LOCALDEBS environment variable, and it will be used
instead of MIRROR when looking for local packages.


Additional targets
==================

Jigdo
-----
You may also want to make the CD images available in jigdo format.
Jigsaw Download, the successor to the Pseudo-Image Kit.
See http://www.debian.org/CD/jigdo-cd/ for more information on jigdo.

Set the DOJIGDO and related variables in CONF.sh. This is no separate target
for jigdo, merely a modification of the "images" targets. You can choose
only .iso generation (default), only .jigdo generation (for highly reduced
disk usage), or both .iso and .jigdo generation.

To generate the jigdo files and templates, you need an
appropriately-patched version of mkisofs/genisoimage, as shipped in
Debian Etch.

The MD5SUMS file generated by the "imagesums" target will contain the MD5
checksums of all generated images, regardless of the DOJIGDO setting. If
no full iso image is available, the MD5sum will be extracted from the
.template file. A note in the Makefile shows how the original file size
can be extracted from the .template in a similar way.

About jigdo "fallback servers":

jigdo works by downloading individual packages and other files from a
normal Debian mirror, and using them to regenerate a CD/DVD image. 
However, the content of Debian mirrors changes over time, files are
added and removed. But jigdo must have access to all files needed for
the image it has to regenerate, even those that have been removed from
the normal Debian mirrors.

A fallback server contains a backup of the Debian FTP space for the
moment the .jigdo files were generated. This backup is made available
under a certain URL which is written to the .jigdo files. jigdo will
*only* revert to the fallback server after an unsuccessful attempt to
retrieve a file from the normal user-selected Debian mirror, so the
bandwidth requirements are modest.

A fallback is even necessary for .jigdo files of the stable release,
because some files (typically documentation or boot floppies) can
change at any time.

debian-cd allows you to automatically create a directory on disc which
is suitable for use as a fallback mirror. It is populated with hard
links to the archive contents. In CONF.sh, simply supply as
JIGDOFALLBACKPATH the name of the directory, and as JIGDOFALLBACKURLS
the URLs under which it will be made available.

About the hook system
=====================

A hook script can be executed at different times during the CD build
process to customise your CDs. You can specify the script by setting
the various HOOK variables in CONF.sh; look there for more information
about what hook points are available.

About the symlink farm
======================

If you don't have write access on the disk where you have the mirror
or if for another reason hardlink cannot be used, you can try to
use a symlink farm. Instead of having real files, your temporary tree
will be filled with symlinks that mkhybrid will change into files when
it will build the image. You'll need to use a special options. You
have 2 lines of options in CONF.sh as examples.

I've never tested the symlink farm ... it may well generate unusable
images. Don't use it for ISO images that will used by many users.

Note that you will also need a patched mkhybrid that does support the
-F option. Have a look here about it :
http://www.chiark.greenend.org.uk/~stevem/DebianCD/

debian-cd's People

Contributors

joeyh avatar rhertzog avatar cjwatson avatar oahong avatar lordsutch avatar kmuto avatar glaubitz avatar vorlonofportland avatar hdeller avatar falk-hueffner avatar aurel32 avatar pkern avatar odyx avatar h01ger avatar sthibaul avatar stevenc99 avatar joshk0 avatar pmhahn avatar

Watchers

 avatar James Cloos avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.