doc/en_US.ISO8859-1/articles/fbsd-from-scratch/article.xml
2013-11-13 07:52:45 +00:00

599 lines
27 KiB
XML

<?xml version="1.0" encoding="iso-8859-1"?>
<!DOCTYPE article PUBLIC "-//FreeBSD//DTD DocBook XML V5.0-Based Extension//EN"
"http://www.FreeBSD.org/XML/share/xml/freebsd50.dtd" [
<!ENTITY scratch.ap "<application xmlns='http://docbook.org/ns/docbook'>FreeBSD From Scratch</application>">
]>
<article xmlns="http://docbook.org/ns/docbook" xmlns:xlink="http://www.w3.org/1999/xlink" version="5.0" xml:lang="en">
<info><title>FreeBSD From Scratch</title>
<author><personname><firstname>Jens</firstname><surname>Schweikhardt</surname></personname><affiliation>
<address><email>schweikh@FreeBSD.org</email></address>
</affiliation></author>
<copyright>
<year>2002,2003,2004,2008</year>
<holder>Jens Schweikhardt</holder>
</copyright>
<legalnotice xml:id="trademarks" role="trademarks">
&tm-attrib.freebsd;
&tm-attrib.adobe;
&tm-attrib.general;
</legalnotice>
<pubdate>$FreeBSD$</pubdate>
<releaseinfo>$FreeBSD$</releaseinfo>
<abstract>
<para>This article describes my efforts at &scratch.ap;: a fully
automated installation of a customized &os; system compiled from
source, including compilation of all your favorite ports and
configured to match your idea of the perfect system. If you
think <command>make world</command> is a wonderful concept,
&scratch.ap; extends it to <command>make evenmore</command>.</para>
</abstract>
</info>
<sect1 xml:id="introduction">
<title>Introduction</title>
<para>Have you ever upgraded your system with <command>make world</command>?
There is a problem if you have only one system on your disks. If
the <buildtarget>installworld</buildtarget> fails partway through,
you are left with a broken system that might not even boot any
longer. Or maybe the <buildtarget>installworld</buildtarget> runs
smoothly but the new kernel does not boot. Then it is time to
reach for the Fixit CD and dig for those backups you have taken
half a year ago.</para>
<para>I believe in the <quote>wipe your disks when upgrading systems</quote>
paradigm. Wiping disks, or rather partitions, makes sure there is no
old cruft left lying around, something which most upgrade procedures
just do not care about. But wiping the partitions means you have to
also recompile/reinstall all your ports and packages and then
redo all your carefully crafted configuration tweaks.
If you think that this task should be automated as well, read on.</para>
</sect1>
<sect1 xml:id="why">
<title>Why would I (not) want &scratch.ap;?</title>
<para>This is a legitimate question. We have
<application>sysinstall</application> and the well known way to
compile the kernel and the userland tools.</para>
<para>The problem with <application>sysinstall</application> is
that it is severely limited in what, where and how it can install.</para>
<itemizedlist>
<listitem>
<para>It is normally used to install pre-built distribution sets and
packages from some other source (CD, DVD, FTP). It cannot install
the result of a <literal>make buildworld</literal>.</para>
</listitem>
<listitem>
<para>It cannot install a second system under a directory
in a running system.</para>
</listitem>
<listitem>
<para>It cannot install in <application>Vinum</application>
or <application>ZFS</application> partitions.</para>
</listitem>
<listitem>
<para>It cannot compile ports, only install precompiled packages.</para>
</listitem>
<listitem>
<para>It is hard to script or to make arbitrary post-installation
changes.</para>
</listitem>
<listitem>
<para>Last but not least, <application>sysinstall</application>
is semi-officially at its End-Of-Life.</para>
</listitem>
</itemizedlist>
<para>The well known way to build and install the world, as
described in <link xlink:href="http://www.FreeBSD.org/doc/en_US.ISO8859-1/books/handbook/makeworld.html">the
Handbook</link>, by default replaces
the existing system. Only the kernel and modules are saved.
System binaries, headers and a lot of other files are overwritten;
obsolete files are still present and can cause surprises. If the
upgrade fails for any reason, it may be hard or even impossible to
restore the previous state of the system.</para>
<para>&scratch.ap; solves all these problems. The strategy is
simple: use a running system to install a new system under an empty
directory tree, while new partitions are mounted appropriately
in that tree. Many config files can be copied to the appropriate
place and &man.mergemaster.8; can take care of those that cannot.
Arbitrary post-configuration of the new system can be
done from within the old system, up to the point where you can
chroot to the new system. In other words, we go through three
stages, where each stage consists of either running a shell
script or invoking <command>make</command>:</para>
<orderedlist>
<listitem>
<para><filename>stage_1.sh</filename>:
Create a new bootable system under an empty directory and merge
or copy as many files as are necessary.
Then boot the new system.</para>
</listitem>
<listitem>
<para><filename>stage_2.sh</filename>:
Install desired ports.</para>
</listitem>
<listitem>
<para><filename>stage_3.mk</filename>:
Do post-configuration for software installed in previous stage.</para>
</listitem>
</orderedlist>
<para>Once you have used &scratch.ap; to build a second system and
found it works satisfactorily for a couple of weeks, you can then
use it again to reinstall the original system. From now on, whenever
you feel like an update is in order, you simply toggle the
partitions you want to wipe and reinstall.</para>
<para>Maybe you have heard of or even tried <link xlink:href="http://www.linuxfromscratch.org/">Linux From Scratch</link>,
or LFS for short. LFS also describes how to build and install a
system from scratch in empty partitions using a running system.
The focus in LFS seems to be to show the role of each system
component (such as kernel, compiler, devices, shell, terminal database,
etc) and the details of each component's installation.
&scratch.ap; does not go into that much detail. My goal is to
provide an automated and complete installation, not explaining all
the gory details that go on under the hood when making the world.
In case you want to explore &os; at this level of detail, start
looking at <filename>/usr/src/Makefile</filename> and follow the
actions of a <command>make buildworld</command>.</para>
<para>There are also downsides in the approach taken by &scratch.ap;
that you should bear in mind.</para>
<!-- XXX: A nice idea would be to write stage_2.sh using a jail
that runs into the newly installed world from stage_1. Having
properly set up a network address as the jail's primary IP
address, it might even be possible to build ports in a chroot
without uninstalling anything from the 'host' system. But
keep in mind that even jails run on the 'host' kernel. -->
<itemizedlist>
<listitem>
<para>While compiling the ports during stage two the system can
not be used for its usual duties. If you run a production server
you have to consider the downtime caused by stage two. The ports
compiled by <filename>stage_2.conf.default</filename> below require
about 8 hours (of which 4 hours are due to
<application>OpenOffice.org</application>) to build on a contemporary
system. If you prefer to install packages instead of ports,
you can significantly reduce the downtime to about 10 minutes.</para>
</listitem>
</itemizedlist>
</sect1>
<sect1 xml:id="prerequisites">
<title>Prerequisites</title>
<para>For going the &scratch.ap; way, you need to have:</para>
<itemizedlist>
<listitem>
<para>A running &os; system with sources and a ports tree.</para>
</listitem>
<listitem>
<para>At least one unused partition where the new system will be
installed.</para>
</listitem>
<listitem>
<para>Experience with running &man.mergemaster.8;. Or at least no fear
doing so.</para>
</listitem>
<listitem>
<para>If you have no or only a slow link to the Internet: the distfiles
for your favorite ports.</para>
</listitem>
<listitem>
<para>Basic knowledge of shell scripting with the Bourne shell,
&man.sh.1;.</para>
</listitem>
<listitem>
<para>Finally, you should also be able to tell your boot
loader how to boot the new system, either interactively, or
by means of a config file.</para>
</listitem>
</itemizedlist>
</sect1>
<sect1 xml:id="stage1">
<title>Stage One: System Installation</title>
<para>The first version of this article used a single shell script
for stage one where all your customization had to be done by editing
the script. After valuable user feedback I have decided to
separate the code and data in the scripts. This allows to have
different configuration data sets to install different systems
without changing any of the code scripts.</para>
<para>The code script for stage one is
<filename>stage_1.sh</filename> and when run with exactly one
argument, like</para>
<informalexample>
<screen>&prompt.root; <userinput>./stage_1.sh default</userinput></screen>
</informalexample>
<para>will read its configuration from
<filename>stage_1.conf.default</filename> and write a log to
<filename>stage_1.log.default</filename>.</para>
<para>Further below you find my <filename>stage_1.conf.default</filename>.
You need to customize it in various places to match your idea of the
<quote>perfect system</quote>. I have tried to extensively comment
the places you should adapt. The configuration script must provide
four shell functions, <command>create_file_systems</command>,
<command>create_etc_fstab</command>, <command>copy_files</command>
and <command>all_remaining_customization</command> (in case it
matters: this is also the sequence in which they will be called
from <filename>stage_1.sh</filename>).</para>
<para>The points to ponder are:</para>
<itemizedlist>
<listitem>
<para>Partition layout.</para>
<para>I do not subscribe to the idea of a single huge partition
for the whole system. My systems generally have at least
one partition for
<filename>/</filename>,
<filename>/usr</filename> and
<filename>/var</filename> with
<filename>/tmp</filename> symlinked to
<filename>/var/tmp</filename>.
In addition I share the file systems for
<filename>/home</filename> (user homes),
<filename>/home/ncvs</filename> (&os; CVS repository replica),
<filename>/usr/ports</filename> (the ports tree),
<filename>/src</filename> (various checked out src trees) and
<filename>/share</filename> (other shared data without the need
for backups, like the news spool).</para>
</listitem>
<listitem>
<para>Luxury items.</para>
<para>What you want immediately after booting the new system and
even before starting stage two. The reason for not simply
chrooting to the new system during stage one and installing
all my beloved ports is that in theory and in practice there
are bootstrap and consistency issues: stage one has your old
kernel running, but the chrooted environment consists of new
binaries and headers. If the new binaries use a new system
call, these binaries will die with <literal>SIGSYS, Bad
system call</literal>, because the old kernel does not have
that system call. I have seen other issues when I tried
building <package>lang/perl5.8</package>.</para>
</listitem>
</itemizedlist>
<para>Before you run <filename>stage_1.sh</filename> make sure
you have completed the usual tasks in preparation for
<command>make installworld installkernel</command>, like:</para>
<itemizedlist>
<listitem>
<para>configured your kernel config file</para>
</listitem>
<listitem>
<para>successfully completed <command>make buildworld</command></para>
</listitem>
<listitem>
<para>successfully completed <command>make buildkernel
KERNCONF=whatever</command></para>
</listitem>
</itemizedlist>
<para>When you run <filename>stage_1.sh</filename> for the first
time, and the config files copied from your running system to the
new system are not up-to-date with respect to what is under
<filename>/usr/src</filename>, <command>mergemaster</command> will
ask you how to proceed. I recommend merging the changes. If you get
tired of going through the dialogues you can simply update the files
on your <emphasis>running</emphasis> system once (Only if this is an
option. You probably do not want to do this if one of your systems
runs <literal>-STABLE</literal> and the other
<literal>-CURRENT</literal>. The changes may be incompatible).
Subsequent <command>mergemaster</command> invocations will detect
that the RCS version IDs match those under
<filename>/usr/src</filename> and skip the file.</para>
<para>The <filename>stage_1.sh</filename> script will stop at the
first command that fails (returns a non-zero exit status) due to
<command>set -e</command>, so you cannot overlook errors. It will
also stop if you use an unset environment variable, probably due
to a typo. You should correct any errors in your version of
<filename>stage_1.conf.default</filename> before you go on.</para>
<para>In <filename>stage_1.sh</filename> we invoke
<command>mergemaster</command>. Even if none of the files requires a
merge, it will display and ask at the end</para>
<screen>*** Comparison complete
*** Saving mtree database for future upgrades
Do you wish to delete what is left of /var/tmp/temproot.stage1? [no] <userinput>no</userinput></screen>
<para>Please answer <literal>no</literal> or just hit
<keycap>Enter</keycap>. The reason is that <command>mergemaster</command>
will have left a few zero sized files below
<filename>/var/tmp/temproot.stage1</filename> which will be copied to the
new system later (unless already there).</para>
<para>After that <command>mergemaster</command> will list the files it
installed and ask if the new <filename>login.conf</filename> should be
generated:</para>
<screen>*** You chose the automatic install option for files that did not
exist on your system. The following were installed for you:
/newroot/etc/defaults/rc.conf
...
/newroot/COPYRIGHT
*** You installed a new aliases file into /newroot/etc/mail, but
the newaliases command is limited to the directories configured
in sendmail.cf. Make sure to create your aliases database by
hand when your sendmail configuration is done.
*** You installed a login.conf file, so make sure that you run
'/usr/bin/cap_mkdb /newroot/etc/login.conf'
to rebuild your login.conf database
Would you like to run it now? y or n [n]</screen>
<para>The answer does not matter since <filename>stage_1.sh</filename> will
run &man.cap.mkdb.1; for you in any case.</para>
<para>Here is the author's <link xlink:href="stage_1.conf.default"><filename>stage_1.conf.default</filename></link>,
which you need to modify substantially. The comments give you
enough information what to change.</para>
<programlisting><xi:include xmlns:xi="http://www.w3.org/2001/XInclude" href="stage_1.conf.default" parse="text"/></programlisting>
<para>Download <link xlink:href="stage_1.conf.default"><filename>stage_1.conf.default</filename>
</link>.</para>
<para>Running this script installs a system that when booted
provides:</para>
<itemizedlist>
<listitem>
<para>Inherited users and groups.</para>
</listitem>
<listitem>
<para>Firewalled Internet connectivity over Ethernet.</para>
</listitem>
<listitem>
<para>Correct time zone and NTP.</para>
</listitem>
<listitem>
<para>Some more minor configuration, like
<filename>/etc/ttys</filename> and
<command>inetd</command>.</para>
</listitem>
</itemizedlist>
<para>Other areas are prepared for configuration, but will not work
until stage two is completed. For example we have copied files to
configure printing and X11. Printing however is likely to need
applications not found in the base system, like &postscript;
utilities. X11 will not run before we have compiled the server,
libraries and programs.</para>
</sect1>
<sect1 xml:id="stage2">
<title>Stage Two: Ports Installation</title>
<note>
<para>It is also possible to install the (precompiled)
packages at this stage, instead of compiling ports. In this case,
<filename>stage_2.sh</filename> would be nothing more than a list of
<command>pkg_add</command> commands. I trust you know how to write
such a script. Here we concentrate on the more flexible and
traditional way of using the ports.</para>
</note>
<para>The following <filename>stage_2.sh</filename> script is how I
install my favorite ports. It can be run any number of times and
will skip all ports that are already installed. It supports the
<emphasis>dryrun</emphasis> option (<option>-n</option>) to just
show what would be done. You run it like <filename>stage_1.sh</filename>
with exactly one argument to denote a config file, e.g.</para>
<informalexample>
<screen>&prompt.root; <userinput>./stage_2.sh default</userinput></screen>
</informalexample>
<para>which will read the list of ports from
<filename>stage_2.conf.default</filename>.</para>
<para>The list of ports consists of lines with two or more space
separated words: the category and the port, optionally followed by
an installation command that will compile and install the port
(default: <command>make install BATCH=yes &lt; /dev/null</command>).
Empty lines and lines
starting with # are ignored. Most of the time it suffices to only
name category and port. A few ports however can be fine tuned by
specifying <command>make</command> variables, e.g.:</para>
<programlisting>www mozilla make WITHOUT_MAILNEWS=yes WITHOUT_CHATZILLA=yes install</programlisting>
<para>In fact you can specify arbitrary shell commands, so you are
not restricted to simple <command>make</command> invocations:</para>
<programlisting>java jdk16 echo true &gt; files/license.sh; make install BATCH=yes &lt; /dev/null
print acroread8 yes accept | make install PAGER=ls
x11-fonts gnu-unifont make install &amp;&amp; mkfontdir /usr/local/lib/X11/fonts/local
news inn-stable CONFIGURE_ARGS="--enable-uucp-rnews --enable-setgid-inews" make install</programlisting>
<para>The first two lines are examples how you can handle ports
asking you to accept a licence. Note how the line for
<package>news/inn-stable</package> is an example
for a one-shot shell variable assignment to
<literal>CONFIGURE_ARGS</literal>. The port
<filename>Makefile</filename> will use this as an initial value
and augment some other essential args. The difference to
specifying a <application>make</application> variable on the command line
with</para>
<programlisting>news inn-stable make CONFIGURE_ARGS="--enable-uucp-rnews --enable-setgid-inews" install</programlisting>
<para>is that the latter will override instead of augment. It depends on
the particular port which method you want.</para>
<para>Be careful that your ports do not use an interactive install, i.e.
they should not try to read from stdin other than what you explicitly
give them on stdin. If they do, they will read the next line(s) from
your list of ports in the here-document and get confused. If
<filename>stage_2.sh</filename> mysteriously skips a port or stops
processing, this is likely the reason.</para>
<para>Below is <filename>stage_2.conf.default</filename>. A log file named
<filename>LOGDIR/category+port</filename> is created for each port
it actually installs.</para>
<programlisting><xi:include xmlns:xi="http://www.w3.org/2001/XInclude" href="stage_2.conf.default" parse="text"/></programlisting>
<para>Download <link xlink:href="stage_2.conf.default"><filename>stage_2.conf.default</filename></link>.</para>
</sect1>
<sect1 xml:id="stage3">
<title>Stage Three</title>
<para>You have installed your beloved ports during stage two. Some
ports require a little bit of configuration. This is what stage three,
the post-configuration is for. I could have integrated this
post-configuration at the end of the <filename>stage_2.sh</filename>
script. However, I think there is a conceptual difference between
installing a port and modifying its out-of-the-box configuration
that warrants a separate stage.</para>
<para>I have chosen to implement stage three as a
<filename>Makefile</filename> because this allows easy selection of
what you want to configure simply by running:</para>
<informalexample>
<screen>&prompt.root; <userinput>make -f stage_3.mk target</userinput></screen>
</informalexample>
<para>As with <filename>stage_2.sh</filename> make sure you have
<filename>stage_3.mk</filename> available after booting the new
system, either by putting it on a shared partition or copying it
somewhere on the new system.</para>
</sect1>
<sect1 xml:id="limitations">
<title>Limitations</title>
<para>The automated installation of a port may prove difficult if it
is interactive and does not support <command>make BATCH=YES
install</command>. For a few ports the interaction is nothing more
than typing <literal>yes</literal> when asked to accept some license.
If the answer is read from standard input, simply pipe the
appropriate answers to the installation command (e.g. <command>yes |
make install</command>. For other ports you need to investigate
where exactly the interactive command is located and deal with it
appropriately. See the examples above for
<package>print/acroread8</package> and
<package>java/jdk16</package>.</para>
<para>You should also be aware of upgrade issues for config files.
In general you do not know when and if the format or contents of a
config file changes. A new group may be added to
<filename>/etc/group</filename>, or <filename>/etc/passwd</filename>
may gain another field. All of this has happened in the past. Simply
copying a config file from the old to the new system may be enough
most of the time, but in these cases it was not. If you update a
system the canonical way (by overwriting the old files) you are
expected to use <command>mergemaster</command> to deal with changes
where you effectively want to merge your local config with
potentially new items. Unfortunately, <command>mergemaster</command>
is only available for base system files, not for anything installed
by ports. Some third party software seems to be especially designed
to keep me on my toes by changing the config file format every
fortnight. To detect such silent changes, I keep a copy of the
modified config files in the same place where I keep
<filename>stage_3.mk</filename> and compare the result with a
<application>make</application> rule, e.g. for
<application>apache</application>'s <filename>httpd.conf</filename>
in target <command>config_apache</command> with</para>
<programlisting>
@if ! cmp -s /usr/local/etc/apache2/httpd.conf httpd.conf; then \
echo "ATTENTION: the httpd.conf has changed. Please examine if"; \
echo "the modifications are still correct. Here is the diff:"; \
diff -u /usr/local/etc/apache2/httpd.conf httpd.conf; \
fi
</programlisting>
<para>If the diff is innocuous I can make the message go away with
<command>cp /usr/local/etc/apache2/httpd.conf
httpd.conf</command>.</para>
<para>I have used &scratch.ap; several times to update a
<literal>7-CURRENT</literal> to <literal>7-CURRENT</literal> and
<literal>8-CURRENT</literal> to <literal>8-CURRENT</literal>, i.e.
I have never tried to install a <literal>8-CURRENT</literal> from
a <literal>7-CURRENT</literal> system or vice versa. Due to the
number of changes between different major release numbers I would
expect this process to be a bit more involved. Using &scratch.ap;
for upgrades within the realm of a <literal>STABLE</literal> branch
should work painlessly (although I have not yet tried it.)</para>
</sect1>
<sect1 xml:id="files">
<title>The Files</title>
<para>Here are the three files you need beside the config files
already shown above.</para>
<para>This is the <link xlink:href="stage_1.sh"><filename>stage_1.sh</filename></link>
script, which you should not need to modify.</para>
<programlisting><xi:include xmlns:xi="http://www.w3.org/2001/XInclude" href="stage_1.sh" parse="text"/></programlisting>
<para>Download <link xlink:href="stage_1.sh"><filename>stage_1.sh</filename></link>.</para>
<para>This is the <link xlink:href="stage_2.sh"><filename>stage_2.sh</filename></link>
script. You may want to modify the variables at the
beginning.</para>
<programlisting><xi:include xmlns:xi="http://www.w3.org/2001/XInclude" href="stage_2.sh" parse="text"/></programlisting>
<para>Download <link xlink:href="stage_2.sh"><filename>stage_2.sh</filename></link>.</para>
<para>This is my <link xlink:href="stage_3.mk"><filename>stage_3.mk</filename></link> to
give you an idea how to automate all reconfiguration.</para>
<programlisting><xi:include xmlns:xi="http://www.w3.org/2001/XInclude" href="stage_3.mk" parse="text"/></programlisting>
<para>Download <link xlink:href="stage_3.mk"><filename>stage_3.mk</filename></link>.</para>
</sect1>
</article>