blog/data/2009-06-24-fasterboot.txt

   1 Title: Debian boots quicker and quicker
   2 Tags: english, debian, bootsystem
   3 Date: 2009-06-24 21:40
   4
   5 <p>I spent Monday and tuesday this week in London with a lot of the
   6 people involved in the boot system on Debian and Ubuntu, to see if we
   7 could find more ways to speed up the boot system.  This was an Ubuntu
   8 funded
   9 <a href="https://wiki.ubuntu.com/FoundationsTeam/BootPerformance/DebianUbuntuSprint">developer
  10 gathering</a>. It was quite productive.  We also discussed the future
  11 of boot systems, and ways to handle the increasing number of boot
  12 issues introduced by the Linux kernel becoming more and more
  13 asynchronous and event base.  The Ubuntu approach using udev and
  14 upstart might be a good way forward.  Time will show.</p>
  15
  16 <p>Anyway, there are a few ways at the moment to speed up the boot
  17 process in Debian.  All of these should be applied to get a quick
  18 boot:</p>
  19
  20 <ul>
  21
  22 <li>Use dash as /bin/sh.</li>
  23
  24 <li>Disable the init.d/hwclock*.sh scripts and make sure the hardware
  25    clock is in UTC.</li>
  26
  27 <li>Install and activate the insserv package to enable
  28    <a href="http://wiki.debian.org/LSBInitScripts/DependencyBasedBoot">dependency
  29    based boot sequencing</a>, and enable concurrent booting.</li>
  30
  31 </ul>
  32
  33 These points are based on the Google summer of code work done by
  34 <a href="http://initscripts-ng.alioth.debian.org/soc2006-bootsystem/">Carlos
  35 Villegas</a>.
  36
  37 <p>Support for makefile-style concurrency during boot was uploaded to
  38 unstable yesterday.  When we tested it, we were able to cut 6 seconds
  39 from the boot sequence.  It depend on very correct dependency
  40 declaration in all init.d scripts, so I expect us to find edge cases
  41 where the dependences in some scripts are slightly wrong when we start
  42 using this.</p>
  43
  44 <p>On our IRC channel for this effort, #pkg-sysvinit, a new idea was
  45 introduced by Raphael Geissert today, one that could affect the
  46 startup speed as well.  Instead of starting some scripts concurrently
  47 from rcS.d/ and another set of scripts from rc2.d/, it would be
  48 possible to run a of them in the same process.  A quick way to test
  49 this would be to enable insserv and run 'mv /etc/rc2.d/S* /etc/rcS.d/;
  50 insserv'.  Will need to test if that work. :)</p>