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

3204 lines
106 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">
<article xmlns="http://docbook.org/ns/docbook" xmlns:xlink="http://www.w3.org/1999/xlink" version="5.0" xml:lang="en">
<info><title>Package Building Procedures</title>
<authorgroup>
<author><orgname>The &os; Ports Management Team</orgname></author>
</authorgroup>
<copyright>
<year>2003</year>
<year>2004</year>
<year>2005</year>
<year>2006</year>
<year>2007</year>
<year>2008</year>
<year>2009</year>
<year>2010</year>
<year>2011</year>
<year>2012</year>
<year>2013</year>
<holder role="mailto:portmgr@FreeBSD.org">The &os; Ports
Management Team</holder>
</copyright>
<legalnotice xml:id="trademarks" role="trademarks">
&tm-attrib.freebsd;
&tm-attrib.intel;
&tm-attrib.sparc;
&tm-attrib.general;
</legalnotice>
<pubdate>$FreeBSD$</pubdate>
<releaseinfo>$FreeBSD$</releaseinfo>
</info>
<sect1 xml:id="intro">
<title>Introduction</title>
<para>In order to provide pre-compiled binaries of third-party
applications for &os;, the Ports Collection is regularly
built on one of the <quote>Package Building Clusters.</quote>
Currently, the main cluster in use is at
<uri xlink:href="http://pointyhat.FreeBSD.org">http://pointyhat.FreeBSD.org</uri>.</para>
<para>This article documents the internal workings of the
cluster.</para>
<note>
<para>Many of the details in this article will be of interest only to
those on the <link xlink:href="&url.base;/portmgr/">Ports Management</link>
team.</para>
</note>
<sect2 xml:id="codebase">
<title>The codebase</title>
<para>Most of the package building magic occurs under the
<filename>/a/portbuild</filename> directory. Unless
otherwise specified, all paths will be relative to
this location. <replaceable>${arch}</replaceable> will
be used to specify one of the package architectures
(e.g., amd64, arm, &i386;, ia64, powerpc, &sparc64;), and
<replaceable>${branch}</replaceable> will be used
to specify the build branch (e.g., 7, 7-exp, 8, 8-exp, 9, 9-exp, 10, 10-exp).
The set of branches that <systemitem class="username">portmgr</systemitem> currently
supports is the same as those that the &os;
<link xlink:href="http://www.freebsd.org/security/index.html#sup">security team</link>
supports.
</para>
<note>
<para>FreeBSD no longer builds packages for branches 4, 5, or 6, nor
for the alpha architecture.</para>
</note>
<para>The scripts that control all of this live in either
<filename>/a/portbuild/scripts/</filename> or.
<filename>/a/portbuild/admin/scripts/</filename>.
These are the checked-out copies from the Subversion repository at
<link xlink:href="http://svnweb.freebsd.org/base/projects/portbuild/">
<filename>base/projects/portbuild/</filename>
</link>.</para>
<para>Typically, incremental builds are done that use previous
packages as dependencies; this takes less time, and puts less
load on the mirrors. Full builds are usually only done:</para>
<itemizedlist>
<listitem>
<para>right after release time, for the
<literal>-STABLE</literal> branches</para>
</listitem>
<listitem>
<para>periodically to test changes to
<literal>-CURRENT</literal></para>
</listitem>
<listitem>
<para>for experimental (<literal>"exp-"</literal>) builds</para>
</listitem>
</itemizedlist>
<para>Packages from experimental builds are not uploaded.</para>
</sect2>
<sect2 xml:id="codebase-notes">
<title>Historical notes on the codebase</title>
<para>Until mid-2010, the scripts were completely specific to
<systemitem>pointyhat.FreeBSD.org</systemitem> as the head (dispatch) node. During
the summer of 2010, a significant rewrite was done in order to allow
for other hosts to be head nodes. Among the changes were:</para>
<itemizedlist>
<listitem>
<para>removal of the hard-coding of the string
<literal>pointyhat</literal></para>
</listitem>
<listitem>
<para>factoring out all configuration constants (which
were previously scattered throughout the code) into configuration
files (see <link linkend="new-head-node">below</link>)</para>
</listitem>
<listitem>
<para>appending the hostname to the directories
specified by <literal>buildid</literal> (this will allow
directories to be unambigious when copied between machines.)</para>
</listitem>
<listitem>
<para>making the scripts more robust in terms of setting
up directories and symlinks</para>
</listitem>
<listitem>
<para>where necessary, changing certain script invocations
to make all the above easier</para>
</listitem>
</itemizedlist>
<note>
<para>Also during this process, the codebase was migrated to the
<link xlink:href="http://svnweb.freebsd.org/base/projects/portbuild/scripts/">
Subversion repository</link>. For reference, the previous version
may still be
<link xlink:href="http://www.freebsd.org/cgi/cvsweb.cgi/ports/Tools/portbuild/scripts/Attic/">
found in CVS</link>.</para>
</note>
</sect2>
<sect2 xml:id="pointyhat-privsep">
<title>Notes on privilege separation</title>
<para>As of January 2013, a rewrite is in progress to further separate
privileges. The following concepts are introduced:</para>
<itemizedlist>
<listitem>
<para>Server-side user <systemitem class="username">portbuild</systemitem> assumes all
responsiblity for operations involving builds and communicating
with the clients. This user no longer has access to
<application>sudo</application>.</para>
</listitem>
<listitem>
<para>Server-side user <systemitem class="username">srcbuild</systemitem> is created
and given responsiblity for operations involving both VCS
operations and anything involving src builds for the clients.
This user does not have access to
<application>sudo</application>.</para>
</listitem>
<listitem>
<para>The server-side
<literal>ports-</literal><replaceable>arch</replaceable>
users go away.</para>
</listitem>
<listitem>
<para>None of the above server-side users have
<application>ssh</application> keys. Individual
<literal>portmgr</literal> will accomplish all those
tasks using <application>ksu</application>. (This is
still work-in-progress.)</para>
</listitem>
<listitem>
<para>The only client-side user is also named
<systemitem class="username">portbuild</systemitem> and still has access to
<application>sudo</application> for the purpose of managing
jails.</para>
</listitem>
</itemizedlist>
</sect2>
</sect1>
<sect1 xml:id="management">
<title>Build Client Management</title>
<para>You may set up clients to either netboot from the master
(<replaceable>connected</replaceable> nodes)
or have them either self-hosted or netboot from some other
<literal>pxe</literal> host
(<replaceable>disconnected</replaceable> nodes).
In all cases they set themselves
up at boot-time to prepare to build packages.</para>
<para>The cluster master <command>rsync</command>s the
interesting data (ports and src trees, bindist tarballs,
scripts, etc.) to disconnected nodes during the node-setup
phase. Then, the disconnected portbuild directory is
nullfs-mounted for jail builds.</para>
<para>The
<systemitem class="username">portbuild</systemitem>
user can &man.ssh.1; to the client nodes to monitor them.
Use <command>sudo</command> and check the
<systemitem>portbuild.<replaceable>hostname</replaceable>.conf</systemitem>
for the user and access details.</para>
<para>The <command>scripts/allgohans</command> script can
be used to run a command on all of the
<replaceable>${arch}</replaceable> clients.</para>
</sect1>
<sect1 xml:id="setup">
<title>Jail Build Environment Setup</title>
<para>Package builds are performed by the clients in a
<literal>jail</literal> populated by the
<filename>portbuild</filename> script using the
<filename>${arch}/${branch}/builds/${buildid}/bindist.tar</filename>
file.</para>
<para>On the server, use the
<command>makeworld</command> command to build a world from the
<filename>${arch}/${branch}/builds/${buildid}/src/</filename>
tree and install it into
<filename>${arch}/${branch}/builds/${buildid}/bindist.tar</filename>.
The tree will
be updated first unless <literal>-novcs</literal> is
specified.</para>
<screen>&prompt.root; <userinput>/a/portbuild/admin/scripts/makeworld ${arch} ${branch} ${buildid} [-novcs]</userinput></screen>
<para>Similiarly on the server, the
<filename>bindist.tar</filename> tarball is created from the
previously installed world by the <command>mkbindist</command>
script.</para>
<screen>&prompt.root; <userinput>/a/portbuild/admin/scripts/mkbindist ${arch} ${branch} ${buildid}</userinput></screen>
<para>The per-machine tarballs are located on the server in
<filename>${arch}/clients</filename>.</para>
<para>The <filename>bindist.tar</filename> file is extracted
onto each client at client boot time, and at the start of
each pass of the <command>dopackages</command>
script.</para>
<para>For both commands above, if
<replaceable>${buildid}</replaceable> is
<literal>latest</literal>, it may be omitted.</para>
<note>
<para>Currently the above two scripts must be run as
<systemitem class="username">root</systemitem>; otherwise, the install scripts
lack sufficient permissions. This is undesirable for
security reasons. Work is in progress in -HEAD to allow
users to do installations; once that is committed, the
intention is to use that and run these two commands as
<systemitem class="username">srcbuild</systemitem>.</para>
</note>
</sect1>
<sect1 xml:id="customizing">
<title>Customizing Your Build</title>
<para>You can customize your build by providing local versions of
<filename>make.conf</filename> and/or
<filename>src.conf</filename>,
named
<filename>${arch}/${branch}/builds/${buildid}/make.conf.server</filename>
and
<filename>${arch}/${branch}/builds/${buildid}/src.conf.server</filename>,
respectively. These will be used in lieu of the default-named
files on the server side.</para>
<para>Similarly, if you wish to also affect the <emphasis>client-side</emphasis>
<filename>make.conf</filename>, you may use
<filename>${arch}/${branch}/builds/${buildid}/make.conf.client</filename>.
</para>
<note>
<para>Due to the fact that individual clients may each have
their own per-host <filename>make.conf</filename>, the
contents of
<filename>${arch}/${branch}/builds/${buildid}/make.conf.client</filename>
will be <emphasis>appended</emphasis> to that
<filename>make.conf</filename>, not supplant it, as is done
for
<filename>${arch}/${branch}/builds/${buildid}/make.conf.server</filename>.</para>
</note>
<note>
<para>There is no similar functionality for
<filename>${arch}/${branch}/builds/${buildid}/src.conf.client</filename>
(what effect would it have?).</para>
</note>
<example>
<title>Sample
<filename>make.conf.target</filename>
to test new default <application>ruby</application>
version</title>
<para>(For this case, the contents are identical for both server
and client.)</para>
<programlisting>RUBY_DEFAULT_VER= 1.9</programlisting>
</example>
<example>
<title>Sample
<filename>make.conf.target</filename>
for <application>clang</application> builds</title>
<para>(For this case, the contents are also identical for both
server and client.)</para>
<programlisting>.if !defined(CC) || ${CC} == "cc"
CC=clang
.endif
.if !defined(CXX) || ${CXX} == "c++"
CXX=clang++
.endif
.if !defined(CPP) || ${CPP} == "cpp"
CPP=clang-cpp
.endif
# Do not die on warnings
NO_WERROR=
WERROR=</programlisting>
</example>
<example>
<title>Sample <filename>make.conf.server</filename> for
<application>pkgng</application></title>
<programlisting>WITH_PKGNG=yes
PKG_BIN=/usr/local/sbin/pkg</programlisting>
</example>
<example>
<title>Sample <filename>make.conf.client</filename> for
<application>pkgng</application></title>
<programlisting>WITH_PKGNG=yes</programlisting>
</example>
<example>
<title>Sample <filename>src.conf.server</filename>
to test new <application>sort</application> codebase</title>
<programlisting>WITH_BSD_SORT=yes</programlisting>
</example>
</sect1>
<sect1 xml:id="starting">
<title>Starting the Build</title>
<para>Separate builds for various combinations of architecture and branch
are supported. All data private to a build (ports tree, src tree,
packages, distfiles, log files, bindist, Makefile, etc) are located under the
<filename>${arch}/${branch}/builds/${buildid}/</filename>
directory.
The most recently created build can be alternatively referenced using buildid
<literal>latest</literal>, and the one before using
<literal>previous</literal>.</para>
<para>New builds are cloned from the <literal>latest</literal>, which is
fast since it uses ZFS.</para>
<sect2 xml:id="build-dopackages">
<title><command>dopackages</command> scripts</title>
<para>The <filename>scripts/dopackages.wrapper</filename> script
is used to perform the builds.</para>
<screen>&prompt.root; <userinput>dopackages.wrapper ${arch} ${branch} ${buildid} [-options]</userinput></screen>
<para>Most often, you will be using <literal>latest</literal> for
the value of <replaceable>buildid</replaceable>.</para>
<para><literal>[-options]</literal> may be zero or more of the
following:</para>
<itemizedlist>
<listitem>
<para><option>-keep</option> - Do not delete this build in the
future, when it would be normally deleted as part of the
<literal>latest</literal> - <literal>previous</literal> cycle.
Do not forget to clean it up manually when you no longer need it.</para>
</listitem>
<listitem>
<para><option>-nofinish</option> - Do not perform
post-processing once the build is complete. Useful
if you expect that the build will need to be restarted
once it finishes. If you use this option, do not forget to cleanup
the clients when you do not need the build any more.</para>
</listitem>
<listitem>
<para><option>-finish</option> - Perform
post-processing only.</para>
</listitem>
<listitem>
<para><option>-nocleanup</option> - By default, when the
<option>-finish</option> stage of the build is complete, the build
data will be deleted from the clients. This option will prevent
that.</para>
</listitem>
<listitem>
<para><option>-restart</option> - Restart an interrupted
(or non-<literal>finish</literal>ed) build from the
beginning. Ports that failed on the previous build will
be rebuilt.</para>
</listitem>
<listitem>
<para><option>-continue</option> - Restart an interrupted
(or non-<literal>finish</literal>ed) build. Will not
rebuild ports that failed on the previous build.</para>
</listitem>
<listitem>
<para><option>-incremental</option> - Compare the
interesting fields of the new
<filename>INDEX</filename> with the previous one,
remove packages and log files for the old ports that
have changed, and rebuild the rest. This
cuts down on build times substantially since
unchanged ports do not get rebuilt every time.</para>
</listitem>
<listitem>
<para><option>-cdrom</option> - This package build is
intended to end up on a CD-ROM, so
<varname>NO_CDROM</varname> packages and distfiles
should be deleted in post-processing.</para>
</listitem>
<listitem>
<para><option>-nobuild</option> - Perform all
the preprocessing steps, but do not actually do
the package build.</para>
</listitem>
<listitem>
<para><option>-noindex</option> - Do not rebuild
<filename>INDEX</filename> during preprocessing.</para>
</listitem>
<listitem>
<para><option>-noduds</option> - Do not rebuild the
<filename>duds</filename> file (ports that are never
built, e.g., those marked <literal>IGNORE</literal>,
<varname>NO_PACKAGE</varname>, etc.) during
preprocessing.</para>
</listitem>
<listitem>
<para><option>-nochecksubdirs</option> - Do not check the
<varname>SUBDIRS</varname> for ports that are not connected
to the build.</para>
</listitem>
<listitem>
<para><option>-trybroken</option> - Try to build
<varname>BROKEN</varname> ports (off by default
because the amd64/&i386; clusters are fast enough now
that when doing incremental builds, more time
was spent rebuilding things that were going to
fail anyway. Conversely, the other clusters
are slow enough that it would be a waste of time
to try and build <varname>BROKEN</varname> ports).</para>
<note>
<para>With <option>-trybroken</option>, you probably
also want to use <option>-fetch-original</option>
and
<option>-unlimited-errors</option>.</para>
</note>
</listitem>
<listitem>
<para><option>-nosrc</option> - Do not update the
<filename>src</filename> tree from the ZFS snapshot, keep the tree from
previous build instead.</para>
</listitem>
<listitem>
<para><option>-srcvcs</option> - Do not update the
<filename>src</filename> tree from the ZFS snapshot, update it with
a fresh checkout instead.</para>
</listitem>
<listitem>
<para><option>-noports</option> - Do not update the
<filename>ports</filename> tree from the ZFS snapshot, keep the tree from
previous build instead.</para>
</listitem>
<listitem>
<para><option>-portsvcs</option> - Do not update the
<filename>ports</filename> tree from the ZFS snapshot, update it with
a fresh checkout instead.</para>
</listitem>
<listitem>
<para><option>-norestr</option> - Do not attempt to build
<varname>RESTRICTED</varname> ports.</para>
</listitem>
<listitem>
<para><option>-noplistcheck</option> - Do not make it fatal for
ports to leave behind files after deinstallation.</para>
</listitem>
<listitem>
<para><option>-nodistfiles</option> - Do not collect distfiles
that pass <command>make checksum</command> for later
uploading to <systemitem>ftp-master</systemitem>.</para>
</listitem>
<listitem>
<para><option>-fetch-original</option> - Fetch the
distfile from the original <varname>MASTER_SITES</varname>
rather than any cache such as on <systemitem>ftp-master</systemitem>.</para>
</listitem>
<listitem>
<para><option>-unlimited-errors</option>
- defeat the "qmanager threshhold" check for runaway
builds. You want this primarily when doing a
<option>-restart</option> of a build that you expect to mostly
fail, or perhaps a <option>-trybroken</option> run. By default,
the threshhold check is done.</para>
</listitem>
</itemizedlist>
<para>Unless you specify <option>-restart</option>,
<option>-continue</option>, or <option>-finish</option>,
the symlinks for the existing builds will be rotated. i.e,
the existing symlink for <filename>previous</filename> will
be deleted; the most recent build will have its symlink changed
to <filename>previous/</filename>; and a new build will be
created and symlinked into <filename>latest/</filename>.</para>
<para>If the last build finished cleanly you do not need to delete
anything. If it was interrupted, or you selected
<option>-nocleanup</option>, you need to clean up clients by running</para>
<screen>&prompt.user; <userinput>build cleanup ${arch} ${branch} ${buildid} -full</userinput></screen>
<para>When a new build is created, the directories <filename>errors/</filename>,
<filename>logs/</filename>, <filename>packages/</filename>, and so
forth, are cleaned by the scripts. If you are short of space,
you can also clean out <filename>ports/distfiles/</filename>.
Leave the <filename>latest/</filename> directory alone; it is
a symlink for the webserver.</para>
<note>
<para><literal>dosetupnodes</literal> is supposed to be run from
the <literal>dopackages</literal> script in the
<option>-restart</option> case, but it can be a good idea to
run it by hand and then verify that the clients all have the
expected job load. Sometimes,
<filename>dosetupnode</filename> cannot clean up a build and you
need to do it by hand. (This is a bug.)</para>
</note>
<para>Make sure the <replaceable>${arch}</replaceable> build
is run as the <systemitem class="username">portbuild</systemitem> user
or it will complain loudly.</para>
<note>
<para>The actual package build itself occurs in two
identical phases. The reason for this is that sometimes
transient problems (e.g., NFS failures, FTP sites being
unreachable, etc.) may halt a build. Doing things
in two phases is a workaround for these types of
problems.</para>
</note>
<para>Be careful that <filename>ports/Makefile</filename>
does not specify any empty subdirectories. This is especially
important if you are doing an -exp build. If the build
process encounters an empty subdirectory, both package build
phases will stop short, and an error similar to the following
will be written to
<filename>${arch}/${branch}/journal</filename>:</para>
<screen>don't know how to make dns-all(continuing)</screen>
<para>To correct this problem, simply comment out or remove
the <varname>SUBDIR</varname> entries that point to empty
subdirectories. After doing this, you can restart the build
by running the proper <command>dopackages</command> command
with the <option>-restart</option> option.</para>
<note>
<para>This problem also appears if you create a new category
<filename>Makefile</filename> with no <varname>SUBDIR</varname>s
in it. This is probably a bug.</para>
</note>
<example>
<title>Update the i386-7 tree and do a complete build</title>
<screen>&prompt.user; <userinput>dopackages.wrapper i386 8 latest -nosrc -norestr -nofinish</userinput></screen>
</example>
<example>
<title>Restart an interrupted amd64-8 build without updating</title>
<screen>&prompt.user; <userinput>dopackages.wrapper amd64 8 latest -nosrc -noports -norestr -continue -noindex -noduds -nofinish</userinput></screen>
</example>
<example>
<title>Post-process a completed sparc64-8 tree</title>
<screen>&prompt.user; <userinput>dopackages.wrapper sparc64 8 -finish</userinput></screen>
</example>
<para>Hint: it is usually best to run the <command>dopackages</command>
command inside of <command>screen(1)</command>.</para>
</sect2>
<sect2 xml:id="build-command">
<title><command>build</command> command</title>
<para>You may need to manipulate the build data before starting it,
especially for experimental builds. This is done with
the <command>build</command> command. Here are the useful
options for creation:</para>
<itemizedlist>
<listitem>
<para><literal>build create arch
branch
[newid]</literal> - Creates
<replaceable>newid</replaceable> (or a datestamp if not specified).
Only needed when bringing up a new branch or a new architecture.</para>
</listitem>
<listitem>
<para><literal>build clone arch
branch oldid
[newid]</literal> - Clones
<replaceable>oldid</replaceable> to
<replaceable>newid</replaceable> (or a datestamp if not specified).</para>
</listitem>
<listitem>
<para><literal>build srcupdate arch
branch
buildid</literal> - Replaces the src
tree with a new ZFS snapshot. Do not forget to use
<literal>-nosrc</literal> flag to <command>dopackages</command>
later!</para>
</listitem>
<listitem>
<para><literal>build portsupdate arch
branch
buildid</literal> - Replaces the ports
tree with a new ZFS snapshot. Do not forget to use
<literal>-noports</literal> flag to <command>dopackages</command>
later!</para>
</listitem>
</itemizedlist>
</sect2>
<sect2 xml:id="build-one">
<title>Building a single package</title>
<para>Sometimes there is a need to rebuild a single package from the
package set. This can be accomplished with the following
invocation:</para>
<screen>&prompt.root; <command>path/qmanager/packagebuild amd64 7-exp 20080904212103 aclock-0.2.3_2.tbz</command></screen>
</sect2>
</sect1>
<sect1 xml:id="anatomy">
<title>Anatomy of a Build</title>
<para>A full build without any <literal>-no</literal>
options performs the following operations in the
specified order:</para>
<orderedlist>
<listitem>
<para>An update of the current <filename>ports</filename>
tree from the ZFS snapshot<footnote xml:id="footnote-status1">
<para>Status of these steps can be found in
<filename>${arch}/${branch}/build.log</filename>
as well as on stderr of the tty running the
<command>dopackages</command> command.</para></footnote></para>
</listitem>
<listitem>
<para>An update of the running branch's
<filename>src</filename> tree from the ZFS snapshot<footnoteref linkend="footnote-status1"/></para>
</listitem>
<listitem>
<para>Checks which ports do not have a
<varname>SUBDIR</varname> entry in their respective
category's <filename>Makefile</filename><footnoteref linkend="footnote-status1"/></para>
</listitem>
<listitem>
<para>Creates the <filename>duds</filename> file, which
is a list of ports not to build<footnoteref linkend="footnote-status1"/><footnote xml:id="footnote-buildstop">
<para>If any of these steps fail, the build will stop
cold in its tracks.</para></footnote></para>
</listitem>
<listitem>
<para>Generates a fresh <filename>INDEX</filename>
file<footnoteref linkend="footnote-status1"/><footnoteref linkend="footnote-buildstop"/></para>
</listitem>
<listitem>
<para>Sets up the nodes that will be used in the
build<footnoteref linkend="footnote-status1"/><footnoteref linkend="footnote-buildstop"/></para>
</listitem>
<listitem>
<para>Builds a list of restricted ports<footnoteref linkend="footnote-status1"/><footnoteref linkend="footnote-buildstop"/></para>
</listitem>
<listitem>
<para>Builds packages (phase 1)<footnote xml:id="footnote-status2"><para>Status of these steps can be found in
<filename>${arch}/${branch}/journal</filename>.
Individual ports will write
their build logs to
<filename>${arch}/${branch}/logs/</filename>
and their error logs to
<filename>${arch}/${branch}/errors/</filename>.
</para></footnote></para>
</listitem>
<listitem>
<para>Performs another node setup<footnoteref linkend="footnote-status1"/></para>
</listitem>
<listitem>
<para>Builds packages (phase 2)<footnoteref linkend="footnote-status2"/></para>
</listitem>
</orderedlist>
</sect1>
<sect1 xml:id="build-maintenance">
<title>Build Maintenance</title>
<para>There are several cases where you will need to manually clean
up a build:</para>
<orderedlist>
<listitem>
<para>You have manually interrupted it.</para>
</listitem>
<listitem>
<para>The head node has been rebooted while
a build was running.</para>
</listitem>
<listitem>
<para><filename>qmanager</filename> has crashed and
has been restarted.</para>
</listitem>
</orderedlist>
<sect2 xml:id="interrupting">
<title>Interrupting a Build</title>
<para>Manually interrupting a build is a bit messy. First you need to
identify the tty in which it's running (either record the output
of &man.tty.1; when you start the build, or use <command>ps x</command>
to identify it. You need to make sure that nothing else important
is running in this tty, e.g., <userinput>ps -t p1</userinput> or whatever.
If there is not, you can just kill off the whole term easily with
<userinput>pkill -t pts/1</userinput>; otherwise issue a
<userinput>kill -HUP</userinput> in there by, for example,
<userinput>ps -t pts/1 -o pid= | xargs kill -HUP</userinput>. Replace
<replaceable>p1</replaceable> by whatever the tty is, of course.</para>
<para>The
package builds dispatched by <command>make</command> to
the client machines will clean themselves up after a
few minutes (check with <command>ps x</command> until they
all go away).</para>
<para>If you do not kill &man.make.1;, then it will spawn more jobs.
If you do not kill <command>dopackages</command>, then it will restart
the entire build. If you do not kill the <command>pdispatch</command>
processes, they'll keep going (or respawn) until they've built their
package.</para>
</sect2>
<sect2 xml:id="cleanup">
<title>Cleaning up a Build</title>
<para>To free up resources, you will need to clean up client machines by
running <command>build cleanup</command> command. For example:</para>
<screen>&prompt.user; <userinput>/a/portbuild/scripts/build cleanup i386 8-exp 20080714120411 -full</userinput></screen>
<para>If you forget to do this, then the old build
<literal>jail</literal>s will not be cleaned up for 24 hours, and no
new jobs will be dispatched in their place since
<systemitem>pointyhat</systemitem> thinks the job slot is still occupied.</para>
<para>To check, <userinput>cat ~/loads/*</userinput> to display the
status of client machines; the first column is the number of jobs
it thinks is running, and this should be roughly concordant
with the load average. <literal>loads</literal> is refreshed
every 2 minutes. If you do <userinput>ps x | grep pdispatch</userinput>
and it is less than the number of jobs that <literal>loads</literal>
thinks are in use, you are in trouble.</para>
<note>
<para>The following notes about mounting only apply to
<literal>connected</literal> nodes.</para>
</note>
<para>You may have problem with the <command>umount</command>
commands hanging. If so, you are going to have to use the
<command>allgohans</command> script to run an &man.ssh.1;
command across all clients for that buildenv. For example:</para>
<screen>&prompt.user; ssh gohan24 df</screen>
<para>will get you a df, and</para>
<screen>&prompt.user; allgohans "umount -f pointyhat.freebsd.org:/var/portbuild/i386/8-exp/ports"
&prompt.user; allgohans "umount -f pointyhat.freebsd.org:/var/portbuild/i386/8-exp/src"</screen>
<para>are supposed to get rid of the hanging mounts. You will have to
keep doing them since there can be multiple mounts.</para>
<note>
<para>Ignore the following:</para>
<screen>umount: pointyhat.freebsd.org:/var/portbuild/i386/8-exp/ports: statfs: No such file or directory
umount: pointyhat.freebsd.org:/var/portbuild/i386/8-exp/ports: unknown file system
umount: Cleanup of /x/tmp/8-exp/chroot/53837/compat/linux/proc failed!
/x/tmp/8-exp/chroot/53837/compat/linux/proc: not a file system root directory</screen>
<para>The former two mean that the client did not have those mounted;
the latter two are a bug.</para>
<para>You may also see messages about <literal>procfs</literal>.</para>
</note>
<note>
<para>The above is the end of the notes that apply only to
<literal>connected</literal> nodes.</para>
</note>
<para>After you have done all the above, remove the
<filename>${arch}/lock</filename>
file before trying to restart the build. If you do not,
<filename>dopackages</filename> will simply exit.</para>
<para>If you have to do a ports tree update before
restarting, you may have to rebuild either <filename>duds</filename>,
<filename>INDEX</filename>, or both.</para>
</sect2>
<sect2 xml:id="build-command-2">
<title>Maintaining builds with the <command>build</command>
command</title>
<para>Here are the rest of the options for the <command>build</command>
command:</para>
<itemizedlist>
<listitem>
<para><literal>build destroy arch
branch</literal> - Destroy the
build id.</para>
</listitem>
<listitem>
<para><literal>build list arch
branch</literal> - Shows the current set
of build ids.</para>
</listitem>
</itemizedlist>
</sect2>
</sect1>
<sect1 xml:id="monitoring">
<title>Monitoring the Build</title>
<para>You can use <command>qclient</command> command to monitor the status
of build nodes, and to list the currently scheduled jobs:</para>
<screen>&prompt.user; <userinput>python path/qmanager/qclient jobs</userinput>
&prompt.user; <userinput>python path/qmanager/qclient status</userinput></screen>
<para>The
<userinput>scripts/stats ${branch}</userinput>
command shows the number of packages already built.</para>
<para>Running <userinput>cat /a/portbuild/*/loads/*</userinput>
shows the client loads and number of concurrent builds in
progress. The files that have been recently updated are the clients
that are online; the others are the offline clients.</para>
<note>
<para>The <command>pdispatch</command> command does the dispatching
of work onto the client, and post-processing.
<command>ptimeout.host</command> is a watchdog that kills a build
after timeouts. So, having 50 <command>pdispatch</command>
processes but only 4 &man.ssh.1; processes means 46
<command>pdispatch</command>es are idle, waiting to get an
idle node.</para>
</note>
<para>Running <userinput>tail -f ${arch}/${branch}/build.log</userinput>
shows the overall build progress.</para>
<para>If a port build is failing, and it is not immediately obvious
from the log as to why, you can preserve the
<varname>WRKDIR</varname> for further analysis. To do this,
touch a file called <filename>.keep</filename> in the port's
directory. The next time the cluster tries to build this port,
it will tar, compress, and copy the <varname>WRKDIR</varname>
to
<filename>${arch}/${branch}/wrkdirs/</filename>.</para>
<para>If you find that the system is looping trying to build the
same package over and over again, you may be able to fix the
problem by rebuilding the offending package by hand.</para>
<para>If all the builds start failing with complaints that they
cannot load the dependent packages, check to see that
<application>httpd</application> is still running, and restart
it if not.</para>
<para>Keep an eye on &man.df.1; output. If the
<filename>/a/portbuild</filename> file system becomes full
then <trademark>Bad Things</trademark> happen.</para>
<para>The status of all current builds is generated periodically
into the <filename>packagestats.html</filename> file, e.g.,
<uri xlink:href="http://pointyhat.FreeBSD.org/errorlogs/packagestats.html">http://pointyhat.FreeBSD.org/errorlogs/packagestats.html</uri>.
For each <literal>buildenv</literal>, the following is displayed:</para>
<itemizedlist>
<listitem>
<para><literal>updated</literal> is the contents of
<filename>.updated</filename>. This is why we recommend that you
update <filename>.updated</filename> for <literal>-exp</literal>
runs (see below).</para>
</listitem>
<listitem>
<para>date of <literal>latest log</literal></para>
</listitem>
<listitem>
<para>number of lines in <filename>INDEX</filename></para>
</listitem>
<listitem>
<para>the number of current <literal>build logs</literal></para>
</listitem>
<listitem>
<para>the number of completed <literal>packages</literal></para>
</listitem>
<listitem>
<para>the number of <literal>errors</literal></para>
</listitem>
<listitem>
<para>the number of duds (shown as <literal>skipped</literal>)</para>
</listitem>
<listitem>
<para><literal>missing</literal> shows the difference between
<filename>INDEX</filename> and the other columns. If you have
restarted a run after a ports tree update, there
will likely be duplicates in the packages and error columns,
and this column will be meaningless. (The script is naive).</para>
</listitem>
<listitem>
<para><literal>running</literal> and <literal>completed</literal>
are guesses based on a &man.grep.1; of <filename>build.log</filename>.
</para>
</listitem>
</itemizedlist>
</sect1>
<sect1 xml:id="errors">
<title>Dealing With Build Errors</title>
<para>The easiest way to track build failures is to receive
the emailed logs and sort them to a folder, so you can maintain a
running list of current failures and detect new ones easily.
To do this, add an email address to
<filename>${branch}/portbuild.conf</filename>.
You can easily bounce the new ones to maintainers.</para>
<para>After a port appears broken on every build combination
multiple times, it is time to mark it <varname>BROKEN</varname>.
Two weeks' notification for the maintainers seems fair.</para>
<note>
<para>To avoid build errors with ports that need to be manually
fetched, put the distfiles into
<filename>~ftp/pub/FreeBSD/distfiles</filename>.
Access restrictions are in place to make sure that only the
build clients can access this directory.</para>
</note>
</sect1>
<sect1 xml:id="release">
<title>Release Builds</title>
<para>When building packages for a release, it may be
necessary to manually update the <literal>ports</literal>
and <filename>src</filename> trees to the release tag and use
<option>-novcs</option> and
<option>-noportsvcs</option>.</para>
<para>To build package sets intended for use on a CD-ROM,
use the <option>-cdrom</option> option to
<command>dopackages</command>.</para>
<para>If the disk space is not available on the cluster, use
<option>-nodistfiles</option> to avoid collecting distfiles.</para>
<para>After the initial build completes, restart the build
with
<option>-restart -fetch-original</option>
to collect updated distfiles as well. Then, once the
build is post-processed, take an inventory of the list
of files fetched:</para>
<screen>&prompt.user; <userinput>cd ${arch}/${branch}</userinput>
&prompt.user; <userinput>find distfiles &gt; distfiles-${release}</userinput></screen>
<para>You should use that output to periodically clean out
the distfiles from <systemitem>ftp-master</systemitem>. When space
gets tight, distfiles from recent releases can be kept while
others can be thrown away.</para>
<para>Once the distfiles have been uploaded (see below),
the final release package set must be created. Just to be
on the safe side, run the
<filename>${arch}/${branch}/cdrom.sh</filename>
script by hand to make sure all the CD-ROM restricted packages
and distfiles have been pruned. Then, copy the
<filename>${arch}/${branch}/packages</filename>
directory to
<filename>${arch}/${branch}/packages-${release}</filename>.
Once the packages are safely moved off, contact the &a.re;
and inform them of the release package location.</para>
<para>Remember to coordinate with the &a.re; about the timing
and status of the release builds.</para>
</sect1>
<sect1 xml:id="uploading">
<title>Uploading Packages</title>
<note>
<para>For FreeBSD.org as of 2013, the instructions
about uploading to <systemitem>ftp-master</systemitem> are obsolete.
In the future, <systemitem>ftp-master</systemitem> will pull
from <systemitem>pointyhat</systemitem>, using a mechanism yet
to be implemented. However, the instructions about
<varname>RESTRICTED</varname> and <varname>NO_CDROM</varname>
must still be <emphasis>carefully</emphasis> followed.</para>
</note>
<para>Once a build has completed, packages and/or distfiles
can be transferred to <systemitem>ftp-master</systemitem> for
propagation to the FTP mirror network. If the build was
run with <option>-nofinish</option>, then make sure to
follow up with
<command>dopackages -finish</command> to post-process the
packages (removes <varname>RESTRICTED</varname> and
<varname>NO_CDROM</varname> packages where appropriate,
prunes packages not listed in <filename>INDEX</filename>,
removes from <filename>INDEX</filename>
references to packages not built, and generates a
<filename>CHECKSUM.MD5</filename>
summary); and distfiles (moves them from the temporary
<filename>distfiles/.pbtmp</filename> directory into
<filename>distfiles/</filename> and removes
<varname>RESTRICTED</varname> and <varname>NO_CDROM</varname>
distfiles).</para>
<para>It is usually a good idea to run the
<command>restricted.sh</command> and/or
<command>cdrom.sh</command> scripts by hand after
<command>dopackages</command> finishes just to be safe.
Run the <command>restricted.sh</command> script before
uploading to <systemitem>ftp-master</systemitem>, then run
<command>cdrom.sh</command> before preparing
the final package set for a release.</para>
<para>The package subdirectories are named by whether they are for
<filename>release</filename>, <filename>stable</filename>, or
<filename>current</filename>. Examples:</para>
<itemizedlist>
<listitem>
<para><filename>packages-7.2-release</filename></para>
</listitem>
<listitem>
<para><filename>packages-7-stable</filename></para>
</listitem>
<listitem>
<para><filename>packages-8-stable</filename></para>
</listitem>
<listitem>
<para><filename>packages-9-stable</filename></para>
</listitem>
<listitem>
<para><filename>packages-10-current</filename></para>
</listitem>
</itemizedlist>
<note>
<para>Some of the directories on
<systemitem>ftp-master</systemitem> are, in fact, symlinks. Examples:</para>
<itemizedlist>
<listitem>
<para><filename>packages-stable</filename></para>
</listitem>
<listitem>
<para><filename>packages-current</filename></para>
</listitem>
</itemizedlist>
<para> Be sure
you move the new packages directory over the
<emphasis>real</emphasis> destination directory, and not
one of the symlinks that points to it.</para>
</note>
<para>If you are doing a completely new package set (e.g., for
a new release), copy packages to the staging area on
<systemitem>ftp-master</systemitem> with something like the following:</para>
<screen>&prompt.root; <userinput>cd /a/portbuild/${arch}/${branch}</userinput>
&prompt.root; <userinput>tar cfv - packages/ | ssh portmgr@ftp-master tar xfC - w/ports/${arch}/tmp/${subdir}</userinput></screen>
<para>Then log into <systemitem>ftp-master</systemitem>, verify that
the package set was transferred successfully, remove the
package set that the new package set is to replace (in
<filename>~/w/ports/${arch}</filename>),
and move the new set into place. (<filename>w/</filename> is
merely a shortcut.)</para>
<para>For incremental builds, packages should be uploaded
using <command>rsync</command> so we do not put too much
strain on the mirrors.</para>
<para><emphasis>ALWAYS</emphasis> use <option>-n</option>
first with <command>rsync</command> and check the output
to make sure it is sane. If it looks good, re-run the
<command>rsync</command> without the <option>-n</option>
option.</para>
<para>Example <command>rsync</command> command for incremental
package upload:</para>
<screen>&prompt.root; <userinput>rsync -n -r -v -l -t -p --delete packages/ portmgr@ftp-master:w/ports/${arch}/${subdir}/ | tee log</userinput></screen>
<para>Distfiles should be transferred with the
<command>cpdistfiles</command> script:</para>
<screen>&prompt.root; <userinput>/a/portbuild/scripts/cpdistfiles ${arch} ${branch} ${buildid} [-yesreally] | tee log2</userinput></screen>
<para>Doing it by hand is deprecated.</para>
</sect1>
<sect1 xml:id="expbuilds">
<title>Experimental Patches Builds</title>
<note>
<para>Most of the information in this section is obsolete
as of 2013 and needs to be rewritten.</para>
</note>
<para>Experimental patches builds are run from time to time to
new features or bugfixes to the ports infrastructure (i.e.
<filename>bsd.port.mk</filename>), or to test large sweeping
upgrades. At any given time there may be several simultaneous
experimental patches branches, such as
<literal>8-exp</literal> on the amd64
architecture.</para>
<para>In general, an experimental patches build is run the same
way as any other build, except that you should first update the
ports tree to the latest version and then apply your patches.
To do the former, you can use the following:</para>
<note>
<para>The following example is obsolete</para>
</note>
<screen>&prompt.user; <userinput>cvs -R update -dP &gt; update.out</userinput>
&prompt.user; <userinput>date &gt; .updated</userinput></screen>
<para>This will most closely simulate what the <literal>dopackages</literal>
script does. (While <filename>.updated</filename> is merely
informative, it can be a help.)</para>
<para>You will need to edit <filename>update.out</filename> to look
for lines beginning with <literal>^M</literal>, <literal>^C</literal>,
or <literal>^?</literal> and then deal with them.</para>
<para>It is always a good idea to save
original copies of all changed files, as well as a list of what
you are changing. You can then look back on this list when doing
the final commit, to make sure you are committing exactly what you
tested.</para>
<para>Since the machine is shared, someone else may delete your
changes by mistake, so keep a copy of them in e.g., your home
directory on <systemitem>freefall</systemitem>. Do not use
<filename>tmp/</filename>; since <systemitem>pointyhat</systemitem>
itself runs some version of <literal>-CURRENT</literal>, you
can expect reboots (if nothing else, for updates).</para>
<para>In order to have a good control case with which to compare
failures, you should first do a package build of the branch on
which the experimental patches branch is based for the &i386;
architecture (currently this is <literal>8</literal>). Then, when
preparing for the experimental patches build, checkout a ports
tree and a src tree with the same date as was used for the control
build. This will ensure an apples-to-apples comparison
later.</para>
<para>Once the build finishes, compare the control build failures
to those of the experimental patches build. Use the following
commands to facilitate this (this assumes the <literal>8</literal>
branch is the control branch, and the <literal>8-exp</literal>
branch is the experimental patches branch):</para>
<screen>&prompt.user; <userinput>cd /a/portbuild/i386/8-exp/errors</userinput>
&prompt.user; <userinput>find . -name \*.log\* | sort &gt; /tmp/8-exp-errs</userinput>
&prompt.user; <userinput>cd /a/portbuild/i386/8/errors</userinput>
&prompt.user; <userinput>find . -name \*.log\* | sort &gt; /tmp/8-errs</userinput></screen>
<note>
<para>If it has been a long time since one of the builds
finished, the logs may have been automatically compressed with
bzip2. In that case, you must use <userinput>sort | sed
's,\.bz2,,g'</userinput> instead.</para>
</note>
<screen>&prompt.user; <userinput>comm -3 /tmp/8-errs /tmp/8-exp-errs | less</userinput></screen>
<para>This last command will produce a two-column report. The
first column is ports that failed on the control build but not in
the experimental patches build; the second column is vice versa.
Reasons that the port might be in the first column
include:</para>
<itemizedlist>
<listitem>
<para>Port was fixed since the control build was run, or was
upgraded to a newer version that is also broken (thus the
newer version should appear in the second column)</para>
</listitem>
<listitem>
<para>Port is fixed by the patches in the experimental patches
build</para>
</listitem>
<listitem>
<para>Port did not build under the experimental patches build
due to a dependency failure</para>
</listitem>
</itemizedlist>
<para>Reasons for a port appearing in the second column
include:</para>
<itemizedlist>
<listitem xml:id="broken-by-exp-patches" xreflabel="broken by experimental patches">
<para>Port was broken by the experimental patches</para>
</listitem>
<listitem xml:id="broken-by-upgrading" xreflabel="broken by upgrading">
<para>Port was upgraded since the control build and has become
broken</para>
</listitem>
<listitem>
<para>Port was broken due to a transient error (e.g., FTP site
down, package client error, etc.)</para>
</listitem>
</itemizedlist>
<para>Both columns should be investigated and the reason for the
errors understood before committing the experimental patches
set. To differentiate between <xref linkend="broken-by-exp-patches"/> and <xref linkend="broken-by-upgrading"/> above, you can do a
rebuild of the affected packages under the control
branch:</para>
<screen>&prompt.user; <userinput>cd /a/portbuild/i386/8/ports</userinput></screen>
<note>
<para>The following example is obsolete</para>
</note>
<note>
<para>Be sure to <userinput>cvs update</userinput> this tree to the same date as
the experimental patches tree.</para>
</note>
<!-- XXX MCL fix -->
<para>The following command will set up the control branch for
the partial build:</para>
<screen>&prompt.user; <userinput>/a/portbuild/scripts/dopackages.wrapper i386 8 latest -noportsvcs -nobuild -novcs -nofinish</userinput></screen>
<!-- XXX MCL obsolete -->
<para>The builds must be performed from the
<filename>packages/All</filename> directory. This directory should
initially be empty except for the Makefile symlink. If this
symlink does not exist, it must be created:</para>
<screen>&prompt.user; <userinput>cd /a/portbuild/i386/8/packages/All</userinput>
&prompt.user; <userinput>ln -sf ../../Makefile .</userinput>
&prompt.user; <userinput>make -k -j&lt;#&gt; &lt;list of packages to build&gt;</userinput></screen>
<note>
<para>&lt;#&gt; is the concurrency of the build to
attempt. It is usually the sum of the weights listed in
<filename>/a/portbuild/i386/mlist</filename> unless you have a
reason to run a heavier or lighter build.</para>
<para>The list of packages to build should be a list of package
names (including versions) as they appear in
<filename>INDEX</filename>. The <varname>PKGSUFFIX</varname>
(i.e., <filename>.tgz</filename> or <filename>.tbz</filename>) is optional.</para>
</note>
<para>This will build only those packages listed as well as all
of their dependencies.</para>
<para>You can check the progress of this
partial build the same way you would a regular build.</para>
<para>Once all
the errors have been resolved, you can commit the package set.
After committing, it is customary to send a <literal>HEADS
UP</literal> email to <link xlink:href="mailto:ports@FreeBSD.org">ports@FreeBSD.org</link> and
copy <link xlink:href="mailto:ports-developers@FreeBSD.org">ports-developers@FreeBSD.org</link>
informing people of the changes. A summary of all changes
should also be committed to
<filename>/usr/ports/CHANGES</filename>.</para>
</sect1>
<sect1 xml:id="new-node">
<title>How to configure a new package building node</title>
<para>Before following these steps, please coordinate with
<literal>portmgr</literal>.</para>
<sect2 xml:id="node-requirements">
<title>Node requirements</title>
<note>
<para>This section is only of interest when considering
tier-2 architectures.</para>
</note>
<para>Here are the requirement for
what a node needs to be generally useful.</para>
<itemizedlist>
<listitem>
<para>CPU capacity: anything less than 500MHz is generally
not useful for package building.</para>
<note>
<para>We are able to adjust the number of jobs
dispatched to each machine, and we generally tune
the number to use 100% of CPU.</para>
</note>
</listitem>
<listitem>
<para>RAM: Less than 2G is not very useful; 8G or more
is preferred. We have been tuning to one job
per 512M of RAM.</para>
</listitem>
<listitem>
<para>disk: at least 20G is needed for filesystem; 32G is
needed for swap. Best performance will be if multiple
disks are used, and configured as <literal>geom</literal>
stripes. Performance numbers are also TBA.</para>
<note>
<para>Package building will test disk drives to destruction.
Be aware of what you are signing up for!</para>
</note>
</listitem>
<listitem>
<para>network bandwidth: TBA. However, an 8-job machine
has been shown to saturate a cable modem line.</para>
</listitem>
</itemizedlist>
</sect2>
<sect2 xml:id="node-preparation">
<title>Preparation</title>
<procedure>
<step>
<para>Pick a unique hostname. It does not have to be
a publicly resolvable hostname (it can be a name on
your internal network).</para>
</step>
<step>
<para>By default, package building requires the following TCP
ports to be accessible: 22 (<literal>ssh</literal>), 414
(<literal>infoseek</literal>), and 8649
(<literal>ganglia</literal>). If these are not accessible,
pick others and ensure that an <command>ssh</command> tunnel
is set up (see below).</para>
<para>(Note: if you have more than one machine at your site,
you will need an individual TCP port for each service on
each machine, and thus <command>ssh</command> tunnels
will be necessary. As such, you will probably need to
configure port forwarding on your firewall.)</para>
</step>
<step>
<para>Decide if you will be booting natively or via
<literal>pxeboot</literal>. You will find that it is
easier to keep up with changes to <literal>-current</literal>
with the latter, especially if you have multiple machines
at your site.</para>
</step>
<step>
<para>Pick a directory to hold ports configuration and
<filename>chroot</filename> subdirectories. It may be
best to put it this on its own partition. (Example:
<filename>/usr2/</filename>.)</para>
<note>
<para>The filename <filename>chroot</filename> is a
historical remnant. The <command>chroot</command>
command is no longer used.</para>
</note>
</step>
<step>
<para>Decide if you will be using a local
<application>squid</application> cache on the client,
instead of the server. It is more efficient to run it
on the server. If you are doing that, skip the "squid"
steps below.)</para>
</step>
</procedure>
</sect2>
<sect2 xml:id="node-src">
<title>Configuring <filename>src</filename></title>
<procedure>
<step>
<para>Create a directory to contain the latest
<literal>-current</literal> source tree and check it
out. (Since your machine will likely be asked to build
packages for <literal>-current</literal>, the kernel it
runs should be reasonably up-to-date with the
<literal>bindist</literal> that will be exported
by our scripts.)</para>
</step>
<step>
<para>If you are using <filename>pxeboot</filename>: create a
directory to contain the install bits. You will probably
want to use a subdirectory of <filename>/pxeroot</filename>,
e.g.,
<filename>/pxeroot/${arch}-${branch}</filename>.
Export that as <varname>DESTDIR</varname>.</para>
</step>
<step>
<para>If you are cross-building, export
<varname>TARGET_ARCH</varname>=<replaceable>${arch}</replaceable>.</para>
<note>
<para>The procedure for cross-building ports is not yet
defined.</para>
</note>
</step>
<step>
<para>Generate a kernel config file. Include
<filename>GENERIC</filename> (or, if on &i386;, and
you are using more than
3.5G, <filename>PAE</filename>).</para>
<para>Required options:</para>
<programlisting>options NULLFS
options TMPFS</programlisting>
<para>Suggested options:</para>
<programlisting>options GEOM_CONCAT
options GEOM_STRIPE
options SHMMAXPGS=65536
options SEMMNI=40
options SEMMNS=240
options SEMUME=40
options SEMMNU=120</programlisting>
<para>If you are interested in debugging general
problems, you may wish to use the following.
However, for unattended operations, it is best
to leave it out:</para>
<programlisting>options ALT_BREAK_TO_DEBUGGER</programlisting>
<para>For <filename>PAE</filename>, it is not currently possible
to load modules. Therefore, if you are running an architecture
that supports Linux emulation, you will need to add:</para>
<programlisting>options COMPAT_LINUX
options LINPROCFS</programlisting>
<para>Also for <filename>PAE</filename>, as of 20110912 you need
the following. This needs to be investigated:</para>
<programlisting>nooption NFSD # New Network Filesystem Server
options NFSCLIENT # Network Filesystem Client
options NFSSERVER # Network Filesystem Server</programlisting>
</step>
<step>
<para>As root, do the usual build steps, e.g.:</para>
<screen>&prompt.root; <userinput>make -j4 buildworld</userinput>
&prompt.root; <userinput>make buildkernel KERNCONF=${kernconf}</userinput>
&prompt.root; <userinput>make installkernel KERNCONF=${kernconf}</userinput>
&prompt.root; <userinput>make installworld</userinput></screen>
<para>The install steps use <varname>DESTDIR</varname>.</para>
</step>
<step>
<para>Customize files in <filename>etc/</filename>.
Whether you do this on the client itself, or another
machine, will depend on whether you are using
<filename>pxeboot</filename>.</para>
<para>If you are using <filename>pxeboot</filename>: create
a subdirectory of
<filename>${DESTDIR}</filename>
called <filename>conf/</filename>. Create one subdirectory
<filename>default/etc/</filename>, and (if your site will host
multiple nodes), subdirectories
<filename>${ip-address}/etc/</filename>
to contain override files for individual hosts. (You may find
it handy to symlink each of those directories to a hostname.)
Copy the entire contents of
<filename>${DESTDIR}/etc/</filename>
to <filename>default/etc/</filename>; that is where you will
edit your files. The by-ip-address
<filename>etc/</filename> directories will probably only need
customized <filename>rc.conf</filename> files.</para>
<para>In either case, apply the following steps:</para>
<itemizedlist>
<listitem>
<para>Create a
<literal>portbuild</literal>
user and group. It can have the <literal>'*'</literal> password.</para>
<para>Create
<filename>/home/portbuild/.ssh/</filename>
and populate <filename>authorized_keys</filename>. </para>
</listitem>
<listitem>
<para>If you are using <application>ganglia</application>
for monitoring, add the following user:</para>
<programlisting>ganglia:*:102:102::0:0:User &amp;:/usr/local/ganglia:/bin/sh</programlisting>
<para>Add it to <filename>etc/group</filename> as well.</para>
</listitem>
<listitem>
<para>If you are using a local <application>squid</application>
cache on the client, add the following user:</para>
<programlisting>squid:*:100:100::0:0:User &amp;:/usr/local/squid:/bin/sh</programlisting>
<para>Add it to <filename>etc/group</filename> as well.</para>
</listitem>
<listitem>
<para>Create the appropriate files in
<filename>etc/.ssh/</filename>.</para>
</listitem>
<listitem>
<para>In <filename>etc/crontab</filename>: add</para>
<programlisting>* * * * * root /var/portbuild/scripts/client-metrics</programlisting>
</listitem>
<listitem>
<para>Create the appropriate
<filename>etc/fstab</filename>. (If you have multiple,
different, machines, you will need to put those in
the override directories.)</para>
</listitem>
<listitem>
<para>In <filename>etc/inetd.conf</filename>: add</para>
<programlisting>infoseek stream tcp nowait nobody /var/portbuild/scripts/reportload</programlisting>
</listitem>
<listitem>
<para>You should run the cluster on UTC. If you have not set the clock
to UTC:</para>
<programlisting>&prompt.root; cp -p /usr/share/zoneinfo/Etc/UTC etc/localtime</programlisting>
</listitem>
<listitem>
<para>Create the appropriate
<filename>etc/rc.conf</filename>. (If you are using
<literal>pxeboot</literal>, and have multiple,
different, machines, you will need to put those in
the override directories.)</para>
<para>Recommended entries for physical nodes:</para>
<programlisting>hostname="<replaceable>${hostname}</replaceable>"
inetd_enable="YES"
linux_enable="YES"
nfs_client_enable="YES"
ntpd_enable="YES"
sendmail_enable="NONE"
sshd_enable="YES"
sshd_program="/usr/local/sbin/sshd"</programlisting>
<para>If you are using <application>ganglia</application>
for monitoring, add the following</para>
<programlisting>gmond_enable="YES"</programlisting>
<para>If you are using a local <application>squid</application>
cache on the client, add the following</para>
<programlisting>squid_enable="YES"
squid_chdir="<filename>/a/squid/logs</filename>"
squid_pidfile="<filename>/a/squid/logs/squid.pid</filename>"</programlisting>
<para>Required entries for VMWare-based nodes:</para>
<programlisting>vmware_guest_vmmemctl_enable="YES"
vmware_guest_guestd_enable="YES"</programlisting>
<para>Recommended entries for VMWare-based nodes:</para>
<programlisting>hostname=""
ifconfig_em0="DHCP"
fsck_y_enable="YES"
inetd_enable="YES"
linux_enable="YES"
nfs_client_enable="YES"
sendmail_enable="NONE"
sshd_enable="YES"
sshd_program="/usr/local/sbin/sshd"
gmond_enable="YES"
squid_enable="YES"
squid_chdir="<filename>/a/squid/logs</filename>"
squid_pidfile="<filename>/a/squid/logs/squid.pid</filename>"</programlisting>
<para>&man.ntpd.8; should <emphasis>not</emphasis>
be enabled for VMWare instances.</para>
<para>Also, it may be possible to leave
<application>squid</application> disabled by default
so as to not have
<filename>/a</filename>
persistent (which should save instantiation time.)
Work is still ongoing.
</para>
</listitem>
<listitem>
<para>Create <filename>etc/resolv.conf</filename>, if
necessary.</para>
</listitem>
<listitem>
<para>Modify <filename>etc/sysctl.conf</filename>:</para>
<screen>9a10,30
&gt; kern.corefile=<filename>/a/%N.core</filename>
&gt; kern.sugid_coredump=1
&gt; #debug.witness_ddb=0
&gt; #debug.witness_watch=0
&gt;
&gt; # squid needs a lot of fds (leak?)
&gt; kern.maxfiles=40000
&gt; kern.maxfilesperproc=30000
&gt;
&gt; # Since the NFS root is static we do not need to check frequently for file changes
&gt; # This saves &gt;75% of NFS traffic
&gt; vfs.nfs.access_cache_timeout=300
&gt; debug.debugger_on_panic=1
&gt;
&gt; # For jailing
&gt; security.jail.sysvipc_allowed=1
&gt; security.jail.allow_raw_sockets=1
&gt; security.jail.chflags_allowed=1
&gt; security.jail.enforce_statfs=1
&gt;
&gt; vfs.lookup_shared=1</screen>
</listitem>
<listitem>
<para>If desired, modify <filename>etc/syslog.conf</filename>
to change the logging destinations to
<literal>@pointyhat.freebsd.org</literal>.</para>
</listitem>
</itemizedlist>
</step>
</procedure>
</sect2>
<sect2 xml:id="node-ports">
<title>Configuring <filename>ports</filename></title>
<procedure>
<step>
<para>Install the following ports:</para>
<programlisting>net/rsync
security/sudo</programlisting>
<para>You may also wish to install:</para>
<programlisting>security/openssh-portable (with HPN on)</programlisting>
<para>If you are using <application>ganglia</application>
for monitoring, install the following:</para>
<programlisting>sysutils/ganglia-monitor-core (with GMETAD off)</programlisting>
<para>If you are using a local <application>squid</application>
cache on the client, install the following</para>
<programlisting>www/squid31 (with SQUID_AUFS on)</programlisting>
</step>
<step>
<para>Customize files in <filename>usr/local/etc/</filename>.
Whether you do this on the client itself, or another
machine, will depend on whether you are using
<filename>pxeboot</filename>.</para>
<note>
<para>The trick of using <filename>conf</filename>
override subdirectories is less effective here, because
you would need to copy over all subdirectories of
<filename>usr/</filename>. This is an implementation
detail of how the pxeboot works.</para>
</note>
<para>Apply the following steps:</para>
<itemizedlist>
<listitem>
<para>If you are using <application>ganglia</application>,
modify
<filename>usr/local/etc/gmond.conf</filename>:</para>
<screen>21,22c21,22
&lt; name = "unspecified"
&lt; owner = "unspecified"
---
&gt; name = "<replaceable>${arch}</replaceable> package build cluster"
&gt; owner = "portmgr@FreeBSD.org"
24c24
&lt; url = "unspecified"
---
&gt; url = "http://pointyhat.freebsd.org"</screen>
<!-- XXX MCL adapted literally from krismail; I do not understand it -->
<para>If there are machines from more than one cluster in the
same multicast domain (basically = LAN) then change the
multicast groups to different values (.71, .72, etc).</para>
</listitem>
<listitem>
<!-- XXX MCL get latest patches from narutos -->
<para>Create
<filename>usr/local/etc/rc.d/portbuild.sh</filename>,
using the appropriate value for
<literal>scratchdir</literal>:</para>
<programlisting>#!/bin/sh
#
# Configure a package build system post-boot
scratchdir=<filename>/a</filename>
ln -sf ${scratchdir}/portbuild /var/
# Identify builds ready for use
cd /var/portbuild/<replaceable>arch</replaceable>
for i in */builds/*; do
if [ -f ${i}/.ready ]; then
mkdir /tmp/.setup-${i##*/}
fi
done
# Flag that we are ready to accept jobs
touch /tmp/.boot_finished</programlisting>
</listitem>
<listitem>
<para>If you are using a local <application>squid</application>
cache, modify,
<filename>usr/local/etc/squid/squid.conf</filename>:</para>
<screen>288,290c288,290
&lt; #auth_param basic children 5
&lt; #auth_param basic realm Squid proxy-caching web server
&lt; #auth_param basic credentialsttl 2 hours
---
&gt; auth_param basic children 5
&gt; auth_param basic realm Squid proxy-caching web server
&gt; auth_param basic credentialsttl 2 hours
611a612
&gt; acl localnet src 127.0.0.0/255.0.0.0
655a657
&gt; http_access allow localnet
2007a2011
&gt; maximum_object_size 400 MB
2828a2838
&gt; negative_ttl 0 minutes</screen>
<para>Also, change <filename>usr/local</filename>
to <filename>usr2</filename> in
<literal>cache_dir</literal>,
<literal>access_log</literal>,
<literal>cache_log</literal>,
<literal>cache_store_log</literal>,
<literal>pid_filename</literal>,
<literal>netdb_filename</literal>,
<literal>coredump_dir</literal>.</para>
<para>Finally, change the <literal>cache_dir</literal>
storage scheme from <literal>ufs</literal> to
<literal>aufs</literal> (offers better performance).</para>
</listitem>
<listitem>
<para>Configure <command>ssh</command>: copy
<filename>etc/ssh</filename> to
<filename>usr/local/etc/ssh</filename> and add
<literal>NoneEnabled yes</literal> to
<filename>sshd_config</filename>.</para>
</listitem>
<listitem>
<note>
<para>This step is under review.</para>
</note>
<para>Create
<filename>usr/local/etc/sudoers/sudoers.d/portbuild</filename>:</para>
<programlisting># local changes for package building
portbuild ALL=(ALL) NOPASSWD: ALL</programlisting>
</listitem>
</itemizedlist>
</step>
</procedure>
</sect2>
<sect2 xml:id="node-configuration">
<title>Configuration on the client itself</title>
<procedure>
<step>
<para>Change into the port/package directory you picked
above, e.g.,
<command>cd /usr2</command>.</para>
</step>
<step>
<para>As root:</para>
<screen>&prompt.root; <userinput>mkdir portbuild</userinput>
&prompt.root; <userinput>chown portbuild:portbuild portbuild</userinput>
&prompt.root; <userinput>mkdir pkgbuild</userinput>
&prompt.root; <userinput>chown portbuild:portbuild pkgbuild</userinput></screen>
<para>If you are using a local <application>squid</application>
cache:</para>
<screen>&prompt.root; <userinput>mkdir squid</userinput>
&prompt.root; <userinput>mkdir squid/cache</userinput>
&prompt.root; <userinput>mkdir squid/logs</userinput>
&prompt.root; <userinput>chown -R squid:squid squid</userinput></screen>
</step>
<!-- XXX MCL adapted literally from krismail; I do not understand it -->
<step>
<para>If clients preserve <filename>/var/portbuild</filename>
between boots then they must either preserve their
<filename>/tmp</filename>, or revalidate their available
builds at boot time (see the script on the <literal>amd64</literal>
machines). They must also clean up stale jails from previous
builds before creating <filename>/tmp/.boot_finished</filename>.</para>
</step>
<step>
<para>Boot the client.</para>
</step>
<step>
<para>If you are using a local <application>squid</application>
cache, as root, initialize the <command>squid</command>
directories:</para>
<screen><userinput>squid -z</userinput></screen>
</step>
</procedure>
</sect2>
<sect2 xml:id="pointyhat-configuration">
<title>Configuration on the server</title>
<para>These steps need to be taken by a <literal>portmgr</literal>
acting as <literal>portbuild</literal>
on the server.</para>
<procedure>
<step>
<para>If any of the default TCP ports is not available (see
above), you will need to create an <command>ssh</command>
tunnel for them and include its invocation command in
<literal>portbuild</literal>'s
<filename>crontab</filename>.</para>
</step>
<step>
<para>Unless you can use the defaults, add an entry to
<filename>/home/portbuild/.ssh/config</filename>
to specify the public IP address, TCP port for
<command>ssh</command>, username, and any other necessary
information.</para>
</step>
<step>
<para>Create
<filename>/a/portbuild/${arch}/clients/bindist-${hostname}.tar</filename>.</para>
<itemizedlist>
<listitem>
<para>Copy one of the existing ones as a template and unpack it
in a temporary directory.</para>
</listitem>
<listitem>
<para>Customize <filename>etc/resolv.conf</filename>
for the local site.</para>
</listitem>
<listitem>
<para>Customize <filename>etc/make.conf</filename> for
FTP fetches for the local site. Note: the nulling-out
of <varname>MASTER_SITE_BACKUP</varname> must be common
to all nodes, but the first entry in
<varname>MASTER_SITE_OVERRIDE</varname> should be the
nearest local FTP mirror. Example:</para>
<programlisting>.if defined(FETCH_ORIGINAL)
MASTER_SITE_BACKUP=
.else
MASTER_SITE_OVERRIDE= \
ftp://<replaceable>friendly-local-ftp-mirror</replaceable>/pub/FreeBSD/ports/distfiles/${DIST_SUBDIR}/ \
ftp://${BACKUP_FTP_SITE}/pub/FreeBSD/distfiles/${DIST_SUBDIR}/
.endif</programlisting>
</listitem>
<listitem>
<para><command>tar</command> it up and move it to the right
location.</para>
</listitem>
</itemizedlist>
<para>Hint: you will need one of these for each machine;
however, if you have multiple machines at one site, you
should create a site-specific one (e.g., in
<filename>/a/portbuild/conf/clients/</filename>)
and symlink to it.</para>
</step>
<step>
<para>Create
<filename>/a/portbuild/${arch}/portbuild-${hostname}</filename>
using one of the existing ones as a guide. This
file contains overrides to
<filename>/a/portbuild/${arch}/portbuild.conf</filename>.</para>
<para>Suggested values:</para>
<programlisting>disconnected=1
scratchdir=<filename>/usr2/pkgbuild</filename>
client_user=portbuild
sudo_cmd="sudo -H"
rsync_gzip=-z
infoseek_host=localhost
infoseek_port=<replaceable>${tunelled-tcp-port}</replaceable></programlisting>
<para>If you will be using <application>squid</application>
on the client:</para>
<programlisting>http_proxy="http://localhost:3128/"
squid_dir=<filename>/usr2/squid</filename></programlisting>
<para>If, instead, you will be using <application>squid</application>
on the server:</para>
<programlisting>
http_proxy="http://<replaceable>servername</replaceable>:3128/"</programlisting>
<para>Possible other values:</para>
<programlisting>use_md_swap=1
md_size=9g
use_zfs=1
scp_cmd="/usr/local/bin/scp"
ssh_cmd="/usr/local/bin/ssh"</programlisting>
</step>
</procedure>
<para>These steps need to be taken by a <literal>portmgr</literal>
acting as <literal>root</literal> on <systemitem>pointyhat</systemitem>.</para>
<procedure>
<step>
<para>Add the public IP address to
<filename>/etc/hosts.allow</filename>. (Remember, multiple
machines can be on the same IP address.)</para>
</step>
<step>
<para>If you are using <application>ganglia</application>,
add an appropriate <literal>data_source</literal> entry to
<filename>/usr/local/etc/gmetad.conf</filename>:</para>
<programlisting>data_source "<replaceable>arch</replaceable>/<replaceable>location</replaceable> Package Build Cluster" 30 <replaceable>hostname</replaceable></programlisting>
<para>You will need to restart <filename>gmetad</filename>.</para>
</step>
</procedure>
</sect2>
<sect2 xml:id="node-enabling">
<title>Enabling the node</title>
<para>These steps need to be taken by a <literal>portmgr</literal>
acting as <literal>portbuild</literal>:</para>
<procedure>
<step>
<para>Ensure that <literal>ssh</literal> to the client
is working by executing
<userinput>ssh hostname uname -a</userinput>.
The actual command is not important; what is important is to
confirm the setup, and also add an entry into
<filename>known_hosts</filename>, once you have confirmed the
node's identity.</para>
</step>
<step>
<para>Populate the client's copy of
<filename>/var/portbuild/scripts/</filename> by something like
<userinput>/a/portbuild/scripts/dosetupnode arch major latest hostname</userinput>.
Verify that you now have files in that directory.</para>
</step>
<step>
<para>Test the other TCP ports by executing
<userinput>telnet hostname portnumber</userinput>.
<literal>414</literal> (or its tunnel) should give you a few lines of status
information including <literal>arch</literal> and
<literal>osversion</literal>; <literal>8649</literal> should
give you an <literal>XML</literal> response from
<literal>ganglia</literal>.</para>
</step>
</procedure>
<para>This step needs to be taken by a <literal>portmgr</literal>
acting as <literal>portbuild</literal>:</para>
<procedure>
<step>
<para>Tell <filename>qmanager</filename> about the node. Example:</para>
<para><userinput>python path/qmanager/qclient add
name=uniquename
arch=arch
osversion=osversion
numcpus=number
haszfs=0
online=1
domain=domain
primarypool=package
pools="package all" maxjobs=1
acl="ports-arch,deny_all"
</userinput></para>
</step>
</procedure>
<para>Finally, again as <literal>portmgr</literal>
acting as <literal>portbuild</literal>:</para>
<procedure>
<step>
<para>Once you are sure that the client is working, tell
<application>pollmachine</application> about it by adding
it to
<filename>/a/portbuild/${arch}/mlist</filename>.</para>
</step>
</procedure>
</sect2>
</sect1>
<sect1 xml:id="new-branch">
<title>How to configure a new &os; branch</title>
<sect2 xml:id="new-branch-pre-qmanager">
<title>Steps necessary before <application>qmanager</application> is started</title>
<para>When a new branch is created, some work needs to
be done to specify that the previous branch is no longer
equivalent to <literal>HEAD</literal>.</para>
<note>
<para>As
<literal>srcbuild</literal>:</para>
</note>
<itemizedlist>
<listitem>
<para>
Edit <filename>/a/portbuild/conf/admin/admin.conf</filename>
with the following changes:</para>
<itemizedlist>
<listitem>
<para>Add <replaceable>new-branch</replaceable> to
<varname>SRC_BRANCHES</varname>.</para>
</listitem>
<listitem>
<para>For what was previously head, change
<varname>SRC_BRANCH_<replaceable>branch</replaceable>_SUBDIR</varname> to
<literal>releng/branch.0</literal>
(literal zero).</para>
</listitem>
<listitem>
<para>Add
<varname>SRC_BRANCH_<replaceable>new-branch</replaceable>_SUBDIR</varname>
<literal>=head</literal>.</para>
</listitem>
</itemizedlist>
</listitem>
<listitem>
<para>Run <command>/a/portbuild/admin/scripts/updatesnap</command> manually.</para>
</listitem>
</itemizedlist>
</sect2>
<sect2 xml:id="new-branch-post-qmanager">
<title>Steps necessary after <application>qmanager</application> is started</title>
<itemizedlist>
<listitem>
<para>For each branch that will be supported, do the following:</para>
<itemizedlist>
<listitem>
<para>As <literal>portbuild</literal>,
kick-start the build for the branch with:</para>
<screen>build create <replaceable>arch</replaceable> <replaceable>branch</replaceable></screen>
</listitem>
<listitem>
<para>As <literal>srcbuild</literal>,
<link linkend="setup">create
<filename>bindist.tar</filename></link>.</para>
</listitem>
</itemizedlist>
</listitem>
</itemizedlist>
</sect2>
</sect1>
<sect1 xml:id="old-branch">
<title>How to delete an unsupported &os; branch</title>
<para>When an old branch goes out of support, there are some
things to garbage-collect.</para>
<itemizedlist>
<listitem>
<para>Edit <filename>/a/portbuild/admin/conf/admin.conf</filename>
with the following changes:</para>
<itemizedlist>
<listitem>
<para>Delete <replaceable>old-branch</replaceable> from
<varname>SRC_BRANCHES</varname>.</para>
</listitem>
<listitem>
<para>Delete
<varname>SRC_BRANCH_<replaceable>old-branch</replaceable>_SUBDIR</varname><literal>=</literal>
<replaceable>whatever</replaceable></para>
</listitem>
</itemizedlist>
</listitem>
<listitem>
<screen>umount a/snap/src-<replaceable>old-branch</replaceable>/src;
umount a/snap/src-<replaceable>old-branch</replaceable>;
zfs destroy -r a/snap/src-<replaceable>old-branch</replaceable></screen>
</listitem>
</itemizedlist>
<itemizedlist>
<listitem>
<para>You will probably find that the following files and
symlinks in <filename>/a/portbuild/errorlogs/</filename>
can be removed:</para>
<itemizedlist>
<listitem>
<para>Files named
<filename>*-old_branch-failure.html</filename></para>
</listitem>
<listitem>
<para>Files named
<filename>buildlogs_*-old_branch-*-logs.txt</filename></para>
</listitem>
<listitem>
<para>Symlinks named
<filename>*-old_branch-previous*</filename></para>
</listitem>
<listitem>
<para>Symlinks named
<filename>*-old_branch-latest*</filename></para>
</listitem>
</itemizedlist>
</listitem>
</itemizedlist>
</sect1>
<sect1 xml:id="rebase-branch">
<title>How to rebase on a supported &os; branch</title>
<para>As of 2011, the philosophy of package building is to build
packages based on <emphasis>the earliest supported release</emphasis>
of each branch. e.g.: if on <literal>RELENG-8</literal>, the
following releases are supported: 8.1, 8.2, 8.3; then
<literal>packages-8-stable</literal> should be built from 8.1.</para>
<para>As releases go End-Of-Life (see
<link xlink:href="http://www.freebsd.org/security/index.html#sup">chart</link>),
a full (not incremental!) package build should be done and uploaded.</para>
<para>The procedure is as follows:</para>
<itemizedlist>
<listitem>
<para>Edit <filename>/a/portbuild/admin/conf/admin.conf</filename>
with the following changes:</para>
<itemizedlist>
<listitem>
<para>Change the value of
<varname>SRC_BRANCH_<replaceable>branch</replaceable>_SUBDIR</varname> to
<literal>releng/</literal><replaceable>branch</replaceable>.<replaceable>N</replaceable>
where <literal>N</literal> is the newest 'oldest' release
for that branch.</para>
</listitem>
</itemizedlist>
</listitem>
<listitem>
<para>Run <command>/a/portbuild/admin/scripts/updatesnap</command> manually.</para>
</listitem>
<listitem>
<para>Run <command>dopackages</command> with <literal>-nobuild</literal>.</para>
</listitem>
<listitem>
<para>Follow the <link linkend="setup">setup procedure</link>.</para>
</listitem>
<listitem>
<para>Now you can run <command>dopackages</command> without <literal>-nobuild</literal>.</para>
</listitem>
</itemizedlist>
</sect1>
<sect1 xml:id="new-arch">
<title>How to configure a new architecture</title>
<sect2 xml:id="new-arch-pre-qmanager">
<title>Steps necessary before <application>qmanager</application> is started</title>
<note>
<para>The next steps are most easily done as user
<literal>portbuild</literal>.</para>
</note>
<note>
<para>The following assumes you have already run
<literal>mkportbuild</literal>.</para>
</note>
<itemizedlist>
<listitem>
<para>As the <literal>portbuild</literal> user, run</para>
<screen>&prompt.user; /a/portbuild/tools/addarch <replaceable>arch</replaceable></screen>
</listitem>
</itemizedlist>
<itemizedlist>
<listitem>
<para>For each branch that will be supported, do the following:</para>
<itemizedlist>
<listitem>
<para>Kick-start the build for the branch with</para>
<screen>&prompt.root; build create <replaceable>arch</replaceable> <replaceable>branch</replaceable></screen>
</listitem>
</itemizedlist>
</listitem>
<listitem>
<para>If you are going to store your historical buildlogs and
errorlogs on your head node's hard drive, you may skip this step.
Otherwise:</para>
<para>Create an external directory and link to it:</para>
<example>
<title>Creating and linking an external archive directory</title>
<screen>&prompt.root; mkdir /dumpster/pointyhat/<replaceable>arch</replaceable>/archive
&prompt.root; ln -s /dumpster/pointyhat/<replaceable>arch</replaceable>/archive archive</screen>
</example>
<note>
<para>(Historical note that only applied to the original
<systemitem>pointyhat.FreeBSD.org</systemitem> installation)</para>
<para>It is possible that <filename>/dumpster/pointyhat</filename>
will not have enough space. In that case, create the archive
directory as
<filename>/dumpster/pointyhat/arch/archive</filename>
and symlink to that.</para>
</note>
</listitem>
<listitem>
<para>Populate <filename>clients</filename> as usual.</para>
</listitem>
<listitem>
<para>Edit <filename>portbuild.conf</filename>
from one of the ones for another architecture.
<literal>addarch</literal> will have created a default
one for you.</para>
</listitem>
<listitem>
<para>Create customized
<filename>portbuild.machinename.conf</filename>
files as appropriate.</para>
</listitem>
<listitem>
<para>If you need to create any tunnels:</para>
<procedure>
<step>
<para>Make a private configuration directory:</para>
<screen>&prompt.root; mkdir /a/portbuild/conf/<replaceable>arch</replaceable></screen>
</step>
<step>
<para>In that directory: create any <filename>dotunnel.*</filename>
scripts needed.</para>
</step>
</procedure>
</listitem>
</itemizedlist>
<note>
<para>As <literal>srcbuild</literal>:</para>
</note>
<itemizedlist>
<listitem>
<para>Add <replaceable>arch</replaceable> to <varname>SUPPORTED_ARCHS</varname> in
<filename>/a/portbuild/admin/conf/admin.conf</filename>.</para>
</listitem>
<listitem>
<para>Add the <replaceable>arch</replaceable> directory to
<filename>/a/portbuild/admin/scripts/zbackup</filename>.
(This is a hack and should go away.)</para>
</listitem>
</itemizedlist>
<itemizedlist>
<listitem>
<para>Enable the appropriate <replaceable>arch</replaceable> entry for
<filename>/a/portbuild/scripts/dologs</filename> to the portbuild
<filename>crontab</filename>. (This is a hack and should go away.)</para>
</listitem>
</itemizedlist>
</sect2>
<sect2 xml:id="new-arch-post-qmanager">
<title>Steps necessary after <application>qmanager</application> is started</title>
<note>
<para>Again as <literal>srcbuild</literal>:</para>
</note>
<itemizedlist>
<listitem>
<para>For each branch that will be supported, do the following:</para>
<itemizedlist>
<listitem>
<para><link linkend="setup">Create
<filename>bindist.tar</filename></link>.</para>
</listitem>
</itemizedlist>
</listitem>
</itemizedlist>
</sect2>
</sect1>
<sect1 xml:id="new-head-node">
<title>How to configure a new head node (pointyhat instance)</title>
<sect2 xml:id="pointyhat-basics">
<title>Basic installation</title>
<procedure>
<step>
<para>Install FreeBSD.</para>
</step>
<step>
<para>Create a user to own the <application>portbuild</application>
repository, such as <literal>portbuild</literal>. It should have the
<literal>'*'</literal> password.</para>
</step>
<step>
<para>Similarly, create a user to own the administration functions
and manage the <application>svn</application>
repositories, such as <literal>srcbuild</literal>. It should have the
<literal>'*'</literal> password.</para>
</step>
<step>
<para>Add the following to <filename>/boot/loader.conf</filename>:</para>
<programlisting>console="vidconsole,comconsole"</programlisting>
</step>
<step>
<para>You should run the cluster on UTC. If you have not set the clock
to UTC:</para>
<programlisting>&prompt.root; cp -p /usr/share/zoneinfo/Etc/UTC /etc/localtime</programlisting>
</step>
<step>
<para>Create the appropriate
<filename>/etc/rc.conf</filename>.</para>
<para>Required entries:</para>
<programlisting>hostname="<replaceable>${hostname}</replaceable>"
sshd_enable="YES"
zfs_enable="YES"</programlisting>
<para>Recommended entries:</para>
<programlisting>background_fsck="NO"
clear_tmp_enable="YES"
dumpdev="AUTO"
fsck_y_enable="YES"
apache22_enable="YES"
apache_flags=""
apache_pidfile="/var/run/httpd.pid"
inetd_enable="YES"
inetd_flags="-l -w"
mountd_enable="YES"
nfs_server_enable="YES"
nfs_server_flags="-u -t -n 12"
nfs_remote_port_only="YES"
ntpd_enable="YES"
rpcbind_enable="YES"
rpc_lockd_enable="NO"
rpc_statd_enable="YES"
sendmail_enable="NONE"
smartd_enable="YES"</programlisting>
<para>If you are using <application>ganglia</application>,
add:</para>
<programlisting>gmetad_enable="YES"
gmond_enable="YES"</programlisting>
<para>If you will be using a <application>squid</application>
cache on the server, rather than the clients:</para>
<programlisting>squid_enable="YES"</programlisting>
</step>
<step>
<para>Create <filename>/etc/resolv.conf</filename>, if
necessary.</para>
</step>
<step>
<para>Create the appropriate files in
<filename>/etc/ssh/</filename>.</para>
</step>
<step>
<para>Add the following to <filename>/etc/sysctl.conf</filename>:</para>
<programlisting>kern.maxfiles=40000
kern.maxfilesperproc=38000
sysctl vfs.usermount=1
sysctl vfs.zfs.super_owner=1</programlisting>
</step>
<step>
<para>Make sure the following change is made to
<filename>/etc/ttys</filename>:</para>
<programlisting>ttyu0 "/usr/libexec/getty std.9600" vt100 on secure</programlisting>
</step>
</procedure>
</sect2>
<sect2 xml:id="pointyhat-src">
<title>Configuring <filename>src</filename></title>
<para>You should be able to install from the most recent release
using only the default kernel configuration.</para>
</sect2>
<sect2 xml:id="pointyhat-ports">
<title>Configuring <filename>ports</filename></title>
<procedure>
<step>
<para>The following ports (or their latest successors) are required:</para>
<programlisting>databases/py-sqlite3
databases/py-sqlalchemy (only SQLITE is needed)
devel/git (WITH_SVN)
devel/py-configobj
devel/py-setuptools
devel/subversion
net/nc
net/rsync
www/apache22 (with EXT_FILTER)</programlisting>
<para>Expect those to bring in, among others:</para>
<programlisting>databases/sqlite3
lang/perl-5.14 (or successor)
lang/python27 (or sucessor)</programlisting>
<para>If you are using <application>ganglia</application>,
add:</para>
<programlisting>sysutils/ganglia-monitor-core (with GMETAD off)
sysutils/ganglia-webfrontend (compile with -DWITHOUT_X11)</programlisting>
<para>If you will be using a <application>squid</application>
cache on the server, rather than the clients:</para>
<programlisting>www/squid (with SQUID_AUFS on)</programlisting>
<para>The following ports (or their latest successors) are strongly suggested:</para>
<programlisting>devel/ccache
mail/postfix
net/isc-dhcp41-server
ports-mgmt/pkg
ports-mgmt/portaudit
ports-mgmt/portmaster
shells/bash
shells/zsh
sysutils/screen</programlisting>
<note>
<para>The use of <application>sudo</application> on the master,
which was formerly required, is
<emphasis>no longer recommended</emphasis>.
</para>
</note>
<para>The following ports (or their latest successors) are handy:</para>
<programlisting>benchmarks/bonnie++
ports-mgmt/pkg_tree
sysutils/dmidecode
sysutils/smartmontools
sysutils/zfs-stats</programlisting>
</step>
</procedure>
</sect2>
<sect2 xml:id="pointyhat-zfs-volume">
<title>Configuring the zfs volume and setting up the repository</title>
<para>The following steps need to be done as euid root.</para>
<para>Here is a quick example:</para>
<example>
<title>The contents of example file <filename>portbuild/tools/example_install</filename></title>
<screen>
#!/bin/sh
#
# example script to drive the "mkportbuild" kickstart file
#
export PORTBUILD_USER=portbuild
export SRCBUILD_USER=srcbuild
export ZFS_VOLUME=a
export ZFS_MOUNTPOINT=/a
export VCS_REPOSITORY=svn://svn0.us-east.FreeBSD.org
#
# create the zpool. the examples here are just suggestions and need to be
# customized for your site.
#
# simple examples:
# zpool create ${ZFS_VOLUME} da1
# zpool create ${ZFS_VOLUME} gprootfs
# more complex example:
# zpool create ${ZFS_VOLUME} mirror da1 da2 mirror da3 da4 mirror da5 da6 mirror da7 da8
#
# check out the kickstart file and run it
#
mkdir -p tmp
svn checkout ${VCS_REPOSITORY}/base/projects/portbuild/admin/tools tmp
sh -x ./tmp/mkportbuild
</screen>
</example>
<para>Here is a detailed explanation of the example:</para>
<procedure>
<step>
<para>Export the value of <varname>PORTBUILD_USER</varname>:</para>
<screen>&prompt.root; export PORTBUILD_USER=<replaceable>portbuild</replaceable></screen>
</step>
<step>
<para>Export the value of <varname>SRCBUILD_USER</varname>:</para>
<screen>&prompt.root; export SRCBUILD_USER=<replaceable>srcbuild</replaceable></screen>
</step>
<step>
<para>Pick a <application>zfs</application> volume name and export
it. We have used <replaceable>a</replaceable> so far to date.</para>
<programlisting>&prompt.root; export ZFS_VOLUME=<replaceable>a</replaceable></programlisting>
</step>
<step>
<para>Pick a mountpoint and export it. We have used
<filename>/a</filename> so far to date.</para>
<screen>&prompt.root; export ZFS_MOUNTPOINT=/<replaceable>a</replaceable></screen>
</step>
<step>
<para>Create the <application>zfs</application> volume
and mount it.</para>
<example>
<title>Creating a <application>zfs</application> volume for portbuild</title>
<screen>&prompt.root; zpool create ${ZFS_VOLUME} mirror da1 da2 mirror da3 da4 mirror da5 da6 mirror da7 da8</screen>
</example>
<note>
<para>The kickstart script defines <application>zfs</application>
<literal>permission sets</literal>, so that the
<replaceable>srcbuild</replaceable> user and
<replaceable>portbuild</replaceable> user may administer
subdirectories of this
volume without having to have root privileges.</para>
</note>
</step>
<step>
<para>Select an <application>svn</application> repository
and export it. See the
<link xlink:href="&url.books.handbook;/svn-mirrors.html">&os; Handbook</link>
for the currently supported list.</para>
<screen>&prompt.root; export VCS_REPOSITORY=<replaceable>svn://svn0.us-east.FreeBSD.org</replaceable></screen>
</step>
<step>
<para>Obtain a copy of the kickstart script into a
temporary directory. (You will not need to keep this
directory later.)</para>
<screen>&prompt.root; mkdir -p /home/<replaceable>portbuild</replaceable>/<replaceable>tmp</replaceable>
&prompt.root; svn checkout ${VCS_REPOSITORY}/base/projects/portbuild/admin/tools /home/<replaceable>portbuild</replaceable>/<replaceable>tmp</replaceable></screen>
</step>
<step>
<para>Run the kickstart script:</para>
<screen>&prompt.root; sh /home/<replaceable>portbuild</replaceable>/<replaceable>tmp</replaceable>/mkportbuild</screen>
<para>This will accomplish all the following steps:</para>
<procedure>
<!-- begin of whitespace-broken area -->
<step>
<para>Create the <filename>portbuild</filename> directory</para>
</step>
<step>
<para>Create and mount a new <application>zfs</application>
filesystem on it</para>
</step>
<step>
<para>Set up the directory</para>
</step>
<step>
<para>Set up the initial repository:</para>
</step>
<!-- end of whitespace-broken area -->
<step>
<para>Set up the <application>zfs</application>
<literal>permission sets</literal>.</para>
</step>
<step>
<para>Split ownerships of subdirectories such that
<replaceable>PORTBUILD_USER</replaceable> owns, and
only owns, files that are used to manage builds and
interact with slaves. The more trustable user
<replaceable>SRCBUILD_USER</replaceable> now owns
everything else.</para>
</step>
</procedure>
</step>
</procedure>
</sect2>
<sect2 xml:id="srcbuild-user-configuration">
<title>Configuring the <application>srcbuild</application>-owned files</title>
<procedure>
<step>
<para>Configure the server by making the following changes to
<filename>/a/portbuild/admin/conf/admin.conf</filename>:</para>
<itemizedlist>
<listitem>
<para>Set <varname>SUPPORTED_ARCHS</varname> to the
list of architectures you wish to build packages for.</para>
</listitem>
<listitem>
<para>For each source branch you will be building for, set
<varname>SRC_BRANCHES</varname> and
<varname>SRC_BRANCH_<replaceable>branch</replaceable>_SUBDIR</varname>
as detailed in <xref linkend="new-branch-pre-qmanager"/>.
You should not need to change
<varname>SRC_BRANCHES_PATTERN</varname>.</para>
</listitem>
<listitem>
<para>Set <varname>ZFS_VOLUME</varname> and
<varname>ZFS_MOUNTPOINT</varname> to whatever you
chose above.</para>
</listitem>
<listitem>
<para>Set <varname>VCS_REPOSITORY</varname> to whatever
you chose above.</para>
</listitem>
<listitem>
<para>Set <varname>MASTER_URL</varname> to the http
URL of your server. This will be stamped into the
package build logs and the indices thereof.</para>
</listitem>
</itemizedlist>
<para>Most of the other default values should be fine.</para>
</step>
</procedure>
</sect2>
<sect2 xml:id="portbuild-user-configuration">
<title>Configuring the <application>portbuild</application>-owned files</title>
<procedure>
<step>
<para>Configure how build slaves will talk to your server
by making the following changes to
<filename>/a/portbuild/conf/client.conf</filename>:</para>
<itemizedlist>
<listitem>
<para>Set <varname>CLIENT_NFS_MASTER</varname> to wherever
your build slaves will PXE boot from. (Possibly, the
hostname of your server.)</para>
</listitem>
<listitem>
<para>Set <varname>CLIENT_BACKUP_FTP_SITE</varname>
to a backup site for FTP fetches; again, possibly
the hostname of your server.</para>
</listitem>
<listitem>
<para>Set <varname>CLIENT_UPLOAD_HOST</varname> to
where completed packages will be uploaded.</para>
</listitem>
</itemizedlist>
<para>Most of the other default values should be fine.</para>
</step>
<step>
<para>Most of the default values in
<filename>/a/portbuild/conf/common.conf</filename>
should be fine. This file holds definitions used by
both the server and all its clients.</para>
</step>
<step>
<para>Configure the server by making the following changes to
<filename>/a/portbuild/conf/server.conf</filename>:</para>
<itemizedlist>
<!--
<listitem>
<para>For each source branch you will be building for, set
<makevar>SRC_BRANCHES</makevar> and
<makevar>SRC_BRANCH_<replaceable>branch</replaceable>_SUBDIR</makevar>
as detailed in <xref linkend="new-branch-pre-qmanager"/>.
You should not need to change
<makevar>SRC_BRANCHES_PATTERN</makevar>.</para>
</listitem>
<listitem>
<para>Set <makevar>ZFS_VOLUME</makevar> and
<makevar>ZFS_MOUNTPOINT</makevar> to whatever you
chose above.</para>
</listitem>
-->
<listitem>
<para>Set <varname>UPLOAD_DIRECTORY</varname>,
<varname>UPLOAD_TARGET</varname>, and
<varname>UPLOAD_USER</varname> as appropriate
for your site.</para>
</listitem>
<!--
<listitem>
<para>Set <makevar>VCS_REPOSITORY</makevar> to whatever
you chose above.</para>
</listitem>
<listitem>
<para>Set <makevar>MASTER_URL</makevar> to the http
URL of your server. This will be stamped into the
package build logs and the indices thereof.</para>
</listitem>
-->
</itemizedlist>
<para>Most of the other default values should be fine.</para>
</step>
</procedure>
</sect2>
<sect2 xml:id="pointyhat-pre-qmanager">
<title>pre-<application>qmanager</application></title>
<procedure>
<step>
<para>For each architecture, follow the steps in
<xref linkend="new-arch-pre-qmanager"/>.</para>
</step>
</procedure>
</sect2>
<sect2 xml:id="pointyhat-qmanager">
<title><application>qmanager</application></title>
<procedure>
<step>
<para>As <literal>root</literal>, copy the following files from
<filename>/a/portbuild/admin/etc/rc.d/</filename> to
<filename>/usr/local/etc/rc.d/</filename>:</para>
<programlisting>pollmachine
qmanager</programlisting>
<para>As root, start each one of them. You may find it handy
to start each under <application>screen</application> for
debugging purposes.</para>
</step>
<step>
<para>Initialize the <application>qmanager</application>
database's acl list:</para>
<note>
<para>This should now be automatically done for you by
the first <command>build</command> command.</para>
</note>
<screen>&prompt.root; python /<replaceable>a</replaceable>/portbuild/qmanager/qclient add_acl name=deny_all uidlist= gidlist= sense=0</screen>
</step>
</procedure>
</sect2>
<sect2 xml:id="pointyhat-src-ports-repos">
<title>Creating src and ports repositories</title>
<procedure>
<step>
<para>As the <replaceable>srcbuild</replaceable> user,
run the following commands manually to create the
<literal>src</literal> and <literal>ports</literal>
repositories, respectively:</para>
<screen>&prompt.user; /<replaceable>a</replaceable>/portbuild/admin/scripts/updatesnap.ports
&prompt.user; /<replaceable>a</replaceable>/portbuild/admin/scripts/updatesnap</screen>
<para>These will be periodically run from the
<replaceable>srcbuild</replaceable>
<filename>crontab</filename>, which you will
install below.</para>
</step>
</procedure>
</sect2>
<sect2 xml:id="pointyhat-other-services">
<title>Other services</title>
<procedure>
<step>
<para>Configure
<filename>/usr/local/etc/apache22/httpd.conf</filename>
as appropriate for your site.</para>
</step>
<step>
<para>Copy <filename>/a/portbuild/admin/conf/apache.conf</filename>
to the appropriate <filename>Includes/</filename> subdirectory, e.g.,
<filename>/usr/local/etc/apache22/Includes/portbuild.conf</filename>.
Configure it as appropriate for your site.</para>
</step>
<step>
<para>Install <filename>/a/portbuild/admin/crontabs/portbuild</filename> as
the <systemitem class="username">portbuild</systemitem> crontab via
<command>crontab -u portbuild -e</command>. If you do
not support all the archs listed there, make sure to comment out
the appropriate <application>dologs</application> entries.</para>
</step>
<step>
<para>Install <filename>/a/portbuild/admin/crontabs/srcbuild</filename> as
the <systemitem class="username">srcbuild</systemitem> crontab via
<command>crontab -u srcbuild -e</command>.</para>
</step>
<step>
<para>If your build slaves will be pxebooted, make sure to
enable the <application>tftp</application> entries in
<filename>/etc/inetd.conf</filename>.</para>
</step>
<step>
<para>Configure mail by doing the following:</para>
<para><command>newaliases</command>.</para>
</step>
</procedure>
</sect2>
<sect2 xml:id="pointyhat-finishing-up">
<title>Finishing up</title>
<procedure>
<step>
<para>For each architecture, follow the steps in
<xref linkend="new-arch-post-qmanager"/>.</para>
</step>
<step>
<para>You will probably find it handy to append
the following to the <varname>PATH</varname> definition for
the <replaceable>portbuild</replaceable> user:</para>
<programlisting>/<replaceable>a</replaceable>/portbuild/scripts:/<replaceable>a</replaceable>/portbuild/tools</programlisting>
</step>
<step>
<para>You will also probably find it handy to append
the following to the <varname>PATH</varname> definition for
the <replaceable>srcbuild</replaceable> user:</para>
<programlisting>/<replaceable>a</replaceable>/portbuild/admin/scripts:/<replaceable>a</replaceable>/portbuild/admin/tools</programlisting>
</step>
</procedure>
<para>You should now be ready to build packages.</para>
</sect2>
</sect1>
<sect1 xml:id="disk-failure">
<title>Procedures for dealing with disk failures</title>
<note>
<para>The following section is particular to <systemitem>freebsd.org</systemitem>
and is somewhat obsolete.</para>
</note>
<para>When a machine has a disk failure (e.g., panics due to read errors,
etc), then we should do the following steps:</para>
<itemizedlist>
<listitem>
<para>Note the time and failure mode (e.g., paste in the
relevant console output) in
<filename>/a/portbuild/${arch}/reboots</filename></para>
</listitem>
<listitem>
<para>For i386 gohan clients, scrub the disk by touching
<filename>/SCRUB</filename> in the nfsroot (e.g.,
<filename>/a/nfs/8.dir1/SCRUB</filename>) and rebooting. This will
<command>dd if=/dev/zero of=/dev/ad0</command> and force the drive to
remap any bad sectors it finds, if it has enough spares left. This is
a temporary measure to extend the lifetime of a drive that is on the
way out.</para>
<note>
<para>For the i386 blade systems another signal of a failing
disk seems to be that the blade will completely hang and be
unresponsive to either console break, or even NMI.</para>
</note>
<para>For other build systems that do not newfs their disk at boot (e.g.,
amd64 systems) this step has to be skipped.</para>
</listitem>
<listitem>
<para>If the problem recurs, then the disk is probably toast.
Take the machine out of <filename>mlist</filename> and (for ata disks)
run <command>smartctl</command> on the drive:</para>
<screen>smartctl -t long /dev/ad0</screen>
<para>It will take about 1/2 hour:</para>
<screen>gohan51# smartctl -t long /dev/ad0
smartctl version 5.38 [i386-portbld-freebsd8.0] Copyright (C) 2002-8
Bruce Allen
Home page is http://smartmontools.sourceforge.net/
=== START OF OFFLINE IMMEDIATE AND SELF-TEST SECTION ===
Sending command: "Execute SMART Extended self-test routine immediately in off-line mode".
Drive command "Execute SMART Extended self-test routine immediately in off-line mode" successful.
Testing has begun.
Please wait 31 minutes for test to complete.
Test will complete after Fri Jul 4 03:59:56 2008
Use smartctl -X to abort test.</screen>
<para>Then <command>smartctl -a /dev/ad0</command> shows the status
after it finishes:</para>
<screen># SMART Self-test log structure revision number 1
# Num Test_Description Status Remaining
LifeTime(hours) LBA_of_first_error
# 1 Extended offline Completed: read failure 80% 15252 319286</screen>
<para>It will also display other data including a log of previous drive
errors. It is possible for the drive to show previous DMA errors
without failing the self-test though (because of sector
remapping).</para>
</listitem>
</itemizedlist>
<para>When a disk has failed, please inform the cluster administrators
so we can try to get it replaced.</para>
</sect1>
</article>