Petter Reinholdtsen

Surprising as it might sound, there are still computers using the traditional Sys V init system, and there probably will be until systemd start working on Hurd and FreeBSD. The upstream project still exist, though, and up until today, the upstream source was available from Savannah via subversion. I am happy to report that this just changed.

The upstream source is now in Git, and consist of three repositories:

• sysvinit
• insserv
• startpar

I do not really spend much time on the project these days, and I has mostly retired, but found it best to migrate the source to a good version control system to help those willing to move it forward.

As usual, if you use Bitcoin and want to show your support of my activities, please send Bitcoin donations to my address 15oWEoG9dUPovwmUL9KWAnYRtNJEkP1u1b.

Yesterday, I tried to unlock a HTC Desire HD phone, and it proved to be a slight challenge. Here is the recipe if I ever need to do it again. It all started by me wanting to try the recipe to set up an hardened Android installation from the Tor project blog on a device I had access to. It is a old mobile phone with a broken microphone The initial idea had been to just install CyanogenMod on it, but did not quite find time to start on it until a few days ago.

The unlock process is supposed to be simple: (1) Boot into the boot loader (press volume down and power at the same time), (2) select 'fastboot' before (3) connecting the device via USB to a Linux machine, (4) request the device identifier token by running 'fastboot oem get_identifier_token', (5) request the device unlocking key using the HTC developer web site and unlock the phone using the key file emailed to you.

Unfortunately, this only work fi you have hboot version 2.00.0029 or newer, and the device I was working on had 2.00.0027. This apparently can be easily fixed by downloading a Windows program and running it on your Windows machine, if you accept the terms Microsoft require you to accept to use Windows - which I do not. So I had to come up with a different approach. I got a lot of help from AndyCap on #nuug, and would not have been able to get this working without him.

First I needed to extract the hboot firmware from the windows binary for HTC Desire HD downloaded as 'the RUU' from HTC. For this there is is a github project named unruu using libunshield. The unshield tool did not recognise the file format, but unruu worked and extracted rom.zip, containing the new hboot firmware and a text file describing which devices it would work for.

Next, I needed to get the new firmware into the device. For this I followed some instructions available from HTC1Guru.com, and ran these commands as root on a Linux machine with Debian testing:

adb reboot-bootloader
fastboot oem rebootRUU
fastboot flash zip rom.zip
fastboot flash zip rom.zip
fastboot reboot

The flash command apparently need to be done twice to take effect, as the first is just preparations and the second one do the flashing. The adb command is just to get to the boot loader menu, so turning the device on while holding volume down and the power button should work too.

With the new hboot version in place I could start following the instructions on the HTC developer web site. I got the device token like this:

fastboot oem get_identifier_token 2>&1 | sed 's/(bootloader) //'

And once I got the unlock code via email, I could use it like this:

fastboot flash unlocktoken Unlock_code.bin

And with that final step in place, the phone was unlocked and I could start stuffing the software of my own choosing into the device. So far I only inserted a replacement recovery image to wipe the phone before I start. We will see what happen next. Perhaps I should install Debian on it. :)

By now, it is well known that Debian Jessie will not be using sysvinit as its boot system by default. But how can one keep using sysvinit in Jessie? It is fairly easy, and here are a few recipes, courtesy of Erich Schubert and Simon McVittie.

If you already are using Wheezy and want to upgrade to Jessie and keep sysvinit as your boot system, create a file /etc/apt/preferences.d/use-sysvinit with this content before you upgrade:

Package: systemd-sysv
Pin: release o=Debian
Pin-Priority: -1

This file content will tell apt and aptitude to not consider installing systemd-sysv as part of any installation and upgrade solution when resolving dependencies, and thus tell it to avoid systemd as a default boot system. The end result should be that the upgraded system keep using sysvinit.

If you are installing Jessie for the first time, there is no way to get sysvinit installed by default (debootstrap used by debian-installer have no option for this), but one can tell the installer to switch to sysvinit before the first boot. Either by using a kernel argument to the installer, or by adding a line to the preseed file used. First, the kernel command line argument:

preseed/late_command="in-target apt-get install --purge -y sysvinit-core"

Next, the line to use in a preseed file:

d-i preseed/late_command string in-target apt-get install -y sysvinit-core

One can of course also do this after the first boot by installing the sysvinit-core package.

I recommend only using sysvinit if you really need it, as the sysvinit boot sequence in Debian have several hardware specific bugs on Linux caused by the fact that it is unpredictable when hardware devices show up during boot. But on the other hand, the new default boot system still have a few rough edges I hope will be fixed before Jessie is released.

Update 2014-11-26: Inspired by a blog post by Torsten Glaser, added --purge to the preseed line.

A few days ago I decided to try to help the Hurd people to get their changes into sysvinit, to allow them to use the normal sysvinit boot system instead of their old one. This follow up on the great Google Summer of Code work done last summer by Justus Winter to get Debian on Hurd working more like Debian on Linux. To get started, I downloaded a prebuilt hard disk image from http://ftp.debian-ports.org/debian-cd/hurd-i386/current/debian-hurd.img.tar.gz, and started it using virt-manager.

The first think I had to do after logging in (root without any password) was to get the network operational. I followed the instructions on the Debian GNU/Hurd ports page and ran these commands as root to get the machine to accept a IP address from the kvm internal DHCP server:

settrans -fgap /dev/netdde /hurd/netdde
kill $(ps -ef|awk '/[p]finet/ { print $2}')
kill $(ps -ef|awk '/[d]evnode/ { print $2}')
dhclient /dev/eth0

After this, the machine had internet connectivity, and I could upgrade it and install the sysvinit packages from experimental and enable it as the default boot system in Hurd.

But before I did that, I set a password on the root user, as ssh is running on the machine it for ssh login to work a password need to be set. Also, note that a bug somewhere in openssh on Hurd block compression from working. Remember to turn that off on the client side.

Run these commands as root to upgrade and test the new sysvinit stuff:

cat > /etc/apt/sources.list.d/experimental.list <<EOF
deb http://http.debian.net/debian/ experimental main
EOF
apt-get update
apt-get dist-upgrade
apt-get install -t experimental initscripts sysv-rc sysvinit \
    sysvinit-core sysvinit-utils
update-alternatives --config runsystem

To reboot after switching boot system, you have to use reboot-hurd instead of just reboot, as there is not yet a sysvinit process able to receive the signals from the normal 'reboot' command. After switching to sysvinit as the boot system, upgrading every package and rebooting, the network come up with DHCP after boot as it should, and the settrans/pkill hack mentioned at the start is no longer needed. But for some strange reason, there are no longer any login prompt in the virtual console, so I logged in using ssh instead.

Note that there are some race conditions in Hurd making the boot fail some times. No idea what the cause is, but hope the Hurd porters figure it out. At least Justus said on IRC (#debian-hurd on irc.debian.org) that they are aware of the problem. A way to reduce the impact is to upgrade to the Hurd packages built by Justus by adding this repository to the machine:

cat > /etc/apt/sources.list.d/hurd-ci.list <<EOF
deb http://darnassus.sceen.net/~teythoon/hurd-ci/ sid main
EOF

At the moment the prebuilt virtual machine get some packages from http://ftp.debian-ports.org/debian, because some of the packages in unstable do not yet include the required patches that are lingering in BTS. This is the completely list of "unofficial" packages installed:

# aptitude search '?narrow(?version(CURRENT),?origin(Debian Ports))'
i   emacs                   - GNU Emacs editor (metapackage)
i   gdb                     - GNU Debugger
i   hurd-recommended        - Miscellaneous translators
i   isc-dhcp-client         - ISC DHCP client
i   isc-dhcp-common         - common files used by all the isc-dhcp* packages
i   libc-bin                - Embedded GNU C Library: Binaries
i   libc-dev-bin            - Embedded GNU C Library: Development binaries
i   libc0.3                 - Embedded GNU C Library: Shared libraries
i A libc0.3-dbg             - Embedded GNU C Library: detached debugging symbols
i   libc0.3-dev             - Embedded GNU C Library: Development Libraries and Hea
i   multiarch-support       - Transitional package to ensure multiarch compatibilit
i A x11-common              - X Window System (X.Org) infrastructure
i   xorg                    - X.Org X Window System
i A xserver-xorg            - X.Org X server
i A xserver-xorg-input-all  - X.Org X server -- input driver metapackage
#

All in all, testing hurd has been an interesting experience. :) X.org did not work out of the box and I never took the time to follow the porters instructions to fix it. This time I was interested in the command line stuff.

If one of the points of switching to a new init system in Debian is to get rid of huge init.d scripts, I doubt we need to switch away from sysvinit and init.d scripts at all. Here is an example init.d script, ie a rewrite of /etc/init.d/rsyslog:

#!/lib/init/init-d-script
### BEGIN INIT INFO
# Provides:          rsyslog
# Required-Start:    $remote_fs $time
# Required-Stop:     umountnfs $time
# X-Stop-After:      sendsigs
# Default-Start:     2 3 4 5
# Default-Stop:      0 1 6
# Short-Description: enhanced syslogd
# Description:       Rsyslog is an enhanced multi-threaded syslogd.
#                    It is quite compatible to stock sysklogd and can be 
#                    used as a drop-in replacement.
### END INIT INFO
DESC="enhanced syslogd"
DAEMON=/usr/sbin/rsyslogd

Pretty minimalistic to me... For the record, the original sysv-rc script was 137 lines, and the above is just 15 lines, most of it meta info/comments.

How to do this, you ask? Well, one create a new script /lib/init/init-d-script looking something like this:

#!/bin/sh

# Define LSB log_* functions.
# Depend on lsb-base (>= 3.2-14) to ensure that this file is present
# and status_of_proc is working.
. /lib/lsb/init-functions

#
# Function that starts the daemon/service

#
do_start()
{
	# Return
	#   0 if daemon has been started
	#   1 if daemon was already running
	#   2 if daemon could not be started
	start-stop-daemon --start --quiet --pidfile $PIDFILE --exec $DAEMON --test > /dev/null \
		|| return 1
	start-stop-daemon --start --quiet --pidfile $PIDFILE --exec $DAEMON -- \
		$DAEMON_ARGS \
		|| return 2
	# Add code here, if necessary, that waits for the process to be ready
	# to handle requests from services started subsequently which depend
	# on this one.  As a last resort, sleep for some time.
}

#
# Function that stops the daemon/service
#
do_stop()
{
	# Return
	#   0 if daemon has been stopped
	#   1 if daemon was already stopped
	#   2 if daemon could not be stopped
	#   other if a failure occurred
	start-stop-daemon --stop --quiet --retry=TERM/30/KILL/5 --pidfile $PIDFILE --name $NAME
	RETVAL="$?"
	[ "$RETVAL" = 2 ] && return 2
	# Wait for children to finish too if this is a daemon that forks
	# and if the daemon is only ever run from this initscript.
	# If the above conditions are not satisfied then add some other code
	# that waits for the process to drop all resources that could be
	# needed by services started subsequently.  A last resort is to
	# sleep for some time.
	start-stop-daemon --stop --quiet --oknodo --retry=0/30/KILL/5 --exec $DAEMON
	[ "$?" = 2 ] && return 2
	# Many daemons don't delete their pidfiles when they exit.
	rm -f $PIDFILE
	return "$RETVAL"
}

#
# Function that sends a SIGHUP to the daemon/service
#
do_reload() {
	#
	# If the daemon can reload its configuration without
	# restarting (for example, when it is sent a SIGHUP),
	# then implement that here.
	#
	start-stop-daemon --stop --signal 1 --quiet --pidfile $PIDFILE --name $NAME
	return 0
}

SCRIPTNAME=$1
scriptbasename="$(basename $1)"
echo "SN: $scriptbasename"
if [ "$scriptbasename" != "init-d-library" ] ; then
    script="$1"
    shift
    . $script
else
    exit 0
fi

NAME=$(basename $DAEMON)
PIDFILE=/var/run/$NAME.pid

# Exit if the package is not installed
#[ -x "$DAEMON" ] || exit 0

# Read configuration variable file if it is present
[ -r /etc/default/$NAME ] && . /etc/default/$NAME

# Load the VERBOSE setting and other rcS variables
. /lib/init/vars.sh

case "$1" in
  start)
	[ "$VERBOSE" != no ] && log_daemon_msg "Starting $DESC" "$NAME"
	do_start
	case "$?" in
		0|1) [ "$VERBOSE" != no ] && log_end_msg 0 ;;
		2) [ "$VERBOSE" != no ] && log_end_msg 1 ;;
	esac
	;;
  stop)
	[ "$VERBOSE" != no ] && log_daemon_msg "Stopping $DESC" "$NAME"
	do_stop
	case "$?" in
		0|1) [ "$VERBOSE" != no ] && log_end_msg 0 ;;
		2) [ "$VERBOSE" != no ] && log_end_msg 1 ;;
	esac
	;;
  status)
	status_of_proc "$DAEMON" "$NAME" && exit 0 || exit $?
	;;
  #reload|force-reload)
	#
	# If do_reload() is not implemented then leave this commented out
	# and leave 'force-reload' as an alias for 'restart'.
	#
	#log_daemon_msg "Reloading $DESC" "$NAME"
	#do_reload
	#log_end_msg $?
	#;;
  restart|force-reload)
	#
	# If the "reload" option is implemented then remove the
	# 'force-reload' alias
	#
	log_daemon_msg "Restarting $DESC" "$NAME"
	do_stop
	case "$?" in
	  0|1)
		do_start
		case "$?" in
			0) log_end_msg 0 ;;
			1) log_end_msg 1 ;; # Old process is still running
			*) log_end_msg 1 ;; # Failed to start
		esac
		;;
	  *)
		# Failed to stop
		log_end_msg 1
		;;
	esac
	;;
  *)
	echo "Usage: $SCRIPTNAME {start|stop|status|restart|force-reload}" >&2
	exit 3
	;;
esac

:

It is based on /etc/init.d/skeleton, and could be improved quite a lot. I did not really polish the approach, so it might not always work out of the box, but you get the idea. I did not try very hard to optimize it nor make it more robust either.

A better argument for switching init system in Debian than reducing the size of init scripts (which is a good thing to do anyway), is to get boot system that is able to handle the kernel events sensibly and robustly, and do not depend on the boot to run sequentially. The boot and the kernel have not behaved sequentially in years.

Wouter Verhelst have some interesting comments and opinions on my blog post on the need to clean up /etc/rcS.d/ in Debian and my blog post about the default KDE desktop in Debian. I only have time to address one small piece of his comment now, and though it best to address the misunderstanding he bring forward:

Currently, a system admin has four options: [...] boot to a single-user system (by adding 'single' to the kernel command line; this runs rcS and rc1 scripts)

This make me believe Wouter believe booting into single user mode and booting into runlevel 1 is the same. I am not surprised he believe this, because it would make sense and is a quite sensible thing to believe. But because the boot in Debian is slightly broken, runlevel 1 do not work properly and it isn't the same as single user mode. I'll try to explain what is actually happing, but it is a bit hard to explain.

Single user mode is defined like this in /etc/inittab: "~~:S:wait:/sbin/sulogin". This means the only thing that is executed in single user mode is sulogin. Single user mode is a boot state "between" the runlevels, and when booting into single user mode, only the scripts in /etc/rcS.d/ are executed before the init process enters the single user state. When switching to runlevel 1, the state is in fact not ending in runlevel 1, but it passes through runlevel 1 and end up in the single user mode (see /etc/rc1.d/S03single, which runs "init -t1 S" to switch to single user mode at the end of runlevel 1. It is confusing that the 'S' (single user) init mode is not the mode enabled by /etc/rcS.d/ (which is more like the initial boot mode).

This summary might make it clearer. When booting for the first time into single user mode, the following commands are executed: "/etc/init.d/rc S; /sbin/sulogin". When booting into runlevel 1, the following commands are executed: "/etc/init.d/rc S; /etc/init.d/rc 1; /sbin/sulogin". A problem show up when trying to continue after visiting single user mode. Not all services are started again as they should, causing the machine to end up in an unpredicatble state. This is why Debian admins recommend rebooting after visiting single user mode.

A similar problem with runlevel 1 is caused by the amount of scripts executed from /etc/rcS.d/. When switching from say runlevel 2 to runlevel 1, the services started from /etc/rcS.d/ are not properly stopped when passing through the scripts in /etc/rc1.d/, and not started again when switching away from runlevel 1 to the runlevels 2-5. I believe the problem is best fixed by moving all the scripts out of /etc/rcS.d/ that are not required to get a functioning single user mode during boot.

I have spent several years investigating the Debian boot system, and discovered this problem a few years ago. I suspect it originates from when sysvinit was introduced into Debian, a long time ago.

In the Debian boot system, several packages include scripts that are started from /etc/rcS.d/. In fact, there is a bite more of them than make sense, and this causes a few problems. What kind of problems, you might ask. There are at least two problems. The first is that it is not possible to recover a machine after switching to runlevel 1. One need to actually reboot to get the machine back to the expected state. The other is that single user boot will sometimes run into problems because some of the subsystems are activated before the root login is presented, causing problems when trying to recover a machine from a problem in that subsystem. A minor additional point is that moving more scripts out of rcS.d/ and into the other rc#.d/ directories will increase the amount of scripts that can run in parallel during boot, and thus decrease the boot time.

So, which scripts should start from rcS.d/. In short, only the scripts that _have_ to execute before the root login prompt is presented during a single user boot should go there. Everything else should go into the numeric runlevels. This means things like lm-sensors, fuse and x11-common should not run from rcS.d, but from the numeric runlevels. Today in Debian, there are around 115 init.d scripts that are started from rcS.d/, and most of them should be moved out. Do your package have one of them? Please help us make single user and runlevel 1 better by moving it.

Scripts setting up the screen, keyboard, system partitions etc. should still be started from rcS.d/, but there is for example no need to have the network enabled before the single user login prompt is presented.

As always, things are not so easy to fix as they sound. To keep Debian systems working while scripts migrate and during upgrades, the scripts need to be moved from rcS.d/ to rc2.d/ in reverse dependency order, ie the scripts that nothing in rcS.d/ depend on can be moved, and the next ones can only be moved when their dependencies have been moved first. This migration must be done sequentially while we ensure that the package system upgrade packages in the right order to keep the system state correct. This will require some coordination when it comes to network related packages, but most of the packages with scripts that should migrate do not have anything in rcS.d/ depending on them. Some packages have already been updated, like the sudo package, while others are still left to do. I wish I had time to work on this myself, but real live constrains make it unlikely that I will find time to push this forward.

The last few days I have done some upgrade testing in Debian, to see if the upgrade from Lenny to Squeeze will go smoothly. A few bugs have been discovered and reported in the process (#585410 in nagios3-cgi, #584879 already fixed in enscript and #584861 in kdebase-workspace-data), and to get a more regular testing going on, I am working on a script to automate the test.

The idea is to create a Lenny chroot and use tasksel to install a Gnome or KDE desktop installation inside the chroot before upgrading it. To ensure no services are started in the chroot, a policy-rc.d script is inserted. To make sure tasksel believe it is to install a desktop on a laptop, the tasksel tests are replaced in the chroot (only acceptable because this is a throw-away chroot).

A naive upgrade from Lenny to Squeeze using aptitude dist-upgrade currently always fail because udev refuses to upgrade with the kernel in Lenny, so to avoid that problem the file /etc/udev/kernel-upgrade is created. The bug report #566000 make me suspect this problem do not trigger in a chroot, but I touch the file anyway to make sure the upgrade go well. Testing on virtual and real hardware have failed me because of udev so far, and creating this file do the trick in such settings anyway. This is a known issue and the current udev behaviour is intended by the udev maintainer because he lack the resources to rewrite udev to keep working with old kernels or something like that. I really wish the udev upstream would keep udev backwards compatible, to avoid such upgrade problem, but given that they fail to do so, I guess documenting the way out of this mess is the best option we got for Debian Squeeze.

Anyway, back to the task at hand, testing upgrades. This test script, which I call upgrade-test for now, is doing the trick:

#!/bin/sh
set -ex

if [ "$1" ] ; then
    desktop=$1
else
    desktop=gnome
fi

from=lenny
to=squeeze

exec < /dev/null
unset LANG
mirror=http://ftp.skolelinux.org/debian
tmpdir=chroot-$from-upgrade-$to-$desktop
fuser -mv .
debootstrap $from $tmpdir $mirror
chroot $tmpdir aptitude update
cat > $tmpdir/usr/sbin/policy-rc.d <<EOF
#!/bin/sh
exit 101
EOF
chmod a+rx $tmpdir/usr/sbin/policy-rc.d
exit_cleanup() {
    umount $tmpdir/proc
}
mount -t proc proc $tmpdir/proc
# Make sure proc is unmounted also on failure
trap exit_cleanup EXIT INT

chroot $tmpdir aptitude -y install debconf-utils

# Make sure tasksel autoselection trigger.  It need the test scripts
# to return the correct answers.
echo tasksel tasksel/desktop multiselect $desktop | \
    chroot $tmpdir debconf-set-selections

# Include the desktop and laptop task
for test in desktop laptop ; do
    echo > $tmpdir/usr/lib/tasksel/tests/$test <<EOF
#!/bin/sh
exit 2
EOF
    chmod a+rx $tmpdir/usr/lib/tasksel/tests/$test
done

DEBIAN_FRONTEND=noninteractive
DEBIAN_PRIORITY=critical
export DEBIAN_FRONTEND DEBIAN_PRIORITY
chroot $tmpdir tasksel --new-install

echo deb $mirror $to main > $tmpdir/etc/apt/sources.list
chroot $tmpdir aptitude update
touch $tmpdir/etc/udev/kernel-upgrade
chroot $tmpdir aptitude -y dist-upgrade
fuser -mv

I suspect it would be useful to test upgrades with both apt-get and with aptitude, but I have not had time to look at how they behave differently so far. I hope to get a cron job running to do the test regularly and post the result on the web. The Gnome upgrade currently work, while the KDE upgrade fail because of the bug in kdebase-workspace-data

I am not quite sure what kind of extract from the huge upgrade logs (KDE 167 KiB, Gnome 516 KiB) it make sense to include in this blog post, so I will refrain from trying. I can report that for Gnome, aptitude report 760 packages upgraded, 448 newly installed, 129 to remove and 1 not upgraded and 1024MB need to be downloaded while for KDE the same numbers are 702 packages upgraded, 507 newly installed, 193 to remove and 0 not upgraded and 1117MB need to be downloaded

I am very happy to notice that the Gnome desktop + laptop upgrade is able to migrate to dependency based boot sequencing and parallel booting without a hitch. Was unsure if there were still bugs with packages failing to clean up their obsolete init.d script during upgrades, and no such problem seem to affect the Gnome desktop+laptop packages.

If Debian is to migrate to upstart on Linux, I expect some init.d scripts to migrate (some of) their operations to upstart job while keeping the init.d for hurd and kfreebsd. The packages with such needs will need a way to get their init.d scripts to behave differently when used with sysvinit and with upstart. Because of this, I had a look at the environment variables set when a init.d script is running under upstart, and when it is not.

With upstart, I notice these environment variables are set when a script is started from rcS.d/ (ignoring some irrelevant ones like COLUMNS):

DEFAULT_RUNLEVEL=2
previous=N
PREVLEVEL=
RUNLEVEL=
runlevel=S
UPSTART_EVENTS=startup
UPSTART_INSTANCE=
UPSTART_JOB=rc-sysinit

With sysvinit, these environment variables are set for the same script.

INIT_VERSION=sysvinit-2.88
previous=N
PREVLEVEL=N
RUNLEVEL=S
runlevel=S

The RUNLEVEL and PREVLEVEL environment variables passed on from sysvinit are not set by upstart. Not sure if it is intentional or not to not be compatible with sysvinit in this regard.

For scripts needing to behave differently when upstart is used, looking for the UPSTART_JOB environment variable seem to be a good choice.

It is strange to watch how a bug in Debian causing KDM to fail to start at boot when an NVidia video card is used is handled. The problem seem to be that the nvidia X.org driver uses a long time to initialize, and this duration is longer than kdm is configured to wait.

I came across two bugs related to this issue, #583312 initially filed against initscripts and passed on to nvidia-glx when it became obvious that the nvidia drivers were involved, and #524751 initially filed against kdm and passed on to src:nvidia-graphics-drivers for unknown reasons.

To me, it seem that no-one is interested in actually solving the problem nvidia video card owners experience and make sure the Debian distribution work out of the box for these users. The nvidia driver maintainers expect kdm to be set up to wait longer, while kdm expect the nvidia driver maintainers to fix the driver to start faster, and while they wait for each other I guess the users end up switching to a distribution that work for them. I have no idea what the solution is, but I am pretty sure that waiting for each other is not it.

I wonder why we end up handling bugs this way.

A few days ago, parallel booting was enabled in Debian/testing. The feature seem to hold up pretty well, but three fairly serious issues are known and should be solved:

• The wicd package seen to break NFS mounting and network setup when parallel booting is enabled. No idea why, but the wicd maintainer seem to be on the case.
• The nvidia X driver seem to have a race condition triggered more easily when parallel booting is in effect. The maintainer is on the case.
• The sysv-rc package fail to properly enable dependency based boot sequencing (the shutdown is broken) when old file-rc users try to switch back to sysv-rc. One way to solve it would be for file-rc to create /etc/init.d/.legacy-bootordering, and another is to try to make sysv-rc more robust. Will investigate some more and probably upload a workaround in sysv-rc to help those trying to move from file-rc to sysv-rc get a working shutdown.

All in all not many surprising issues, and all of them seem solvable before Squeeze is released. In addition to these there are some packages with bugs in their dependencies and run level settings, which I expect will be fixed in a reasonable time span.

If you report any problems with dependencies in init.d scripts to the BTS, please usertag the report to get it to show up at the list of usertagged bugs related to this.

Update: Correct bug number to file-rc issue.

Since this evening, parallel booting is the default in Debian/unstable for machines using dependency based boot sequencing. Apparently the testing of concurrent booting has been wider than expected, if I am to believe the input on debian-devel@, and I concluded a few days ago to move forward with the feature this weekend, to give us some time to detect any remaining problems before Squeeze is frozen. If serious problems are detected, it is simple to change the default back to sequential boot. The upload of the new sysvinit package also activate a new upstream version.

More information about dependency based boot sequencing is available from the Debian wiki. It is currently possible to disable parallel booting when one run into problems caused by it, by adding this line to /etc/default/rcS:

CONCURRENCY=none

If you report any problems with dependencies in init.d scripts to the BTS, please usertag the report to get it to show up at the list of usertagged bugs related to this.

The last few days a new boot system called systemd has been introduced to the free software world. I have not yet had time to play around with it, but it seem to be a very interesting alternative to upstart, and might prove to be a good alternative for Debian when we are able to switch to an event based boot system. Tollef is in the process of getting systemd into Debian, and I look forward to seeing how well it work. I like the fact that systemd handles init.d scripts with dependency information natively, allowing them to run in parallel where upstart at the moment do not.

Unfortunately do systemd have the same problem as upstart regarding platform support. It only work on recent Linux kernels, and also need some new kernel features enabled to function properly. This means kFreeBSD and Hurd ports of Debian will need a port or a different boot system. Not sure how that will be handled if systemd proves to be the way forward.

In the mean time, based on the input on debian-devel@ regarding parallel booting in Debian, I have decided to enable full parallel booting as the default in Debian as soon as possible (probably this weekend or early next week), to see if there are any remaining serious bugs in the init.d dependencies. A new version of the sysvinit package implementing this change is already in experimental. If all go well, Squeeze will be released with parallel booting enabled by default.

These days, the init.d script dependencies in Squeeze are quite complete, so complete that it is actually possible to run all the init.d scripts in parallell based on these dependencies. If you want to test your Squeeze system, make sure dependency based boot sequencing is enabled, and add this line to /etc/default/rcS:

CONCURRENCY=makefile

That is it. It will cause sysv-rc to use the startpar tool to run scripts in parallel using the dependency information stored in /etc/init.d/.depend.boot, /etc/init.d/.depend.start and /etc/init.d/.depend.stop to order the scripts. Startpar is configured to try to start the kdm and gdm scripts as early as possible, and will start the facilities required by kdm or gdm as early as possible to make this happen.

Give it a try, and see if you like the result. If some services fail to start properly, it is most likely because they have incomplete init.d script dependencies in their startup script (or some of their dependent scripts have incomplete dependencies). Report bugs and get the package maintainers to fix it. :)

Running scripts in parallel could be the default in Debian when we manage to get the init.d script dependencies complete and correct. I expect we will get there in Squeeze+1, if we get manage to test and fix the remaining issues.

If you report any problems with dependencies in init.d scripts to the BTS, please usertag the report to get it to show up at the list of usertagged bugs related to this.

Since this evening, with the upload of sysvinit version 2.87dsf-2, and the upload of insserv version 1.12.0-10 yesterday, Debian unstable have been migrated to using dependency based boot sequencing. This conclude work me and others have been doing for the last three days. It feels great to see this finally part of the default Debian installation. Now we just need to weed out the last few problems that are bound to show up, to get everything ready for Squeeze.

The next step is migrating /sbin/init from sysvinit to upstart, and fixing the more fundamental problem of handing the event based non-predictable kernel in the early boot.

After several years of frustration with the lack of activity from the existing sysvinit upstream developer, I decided a few weeks ago to take over the package and become the new upstream. The number of patches to track for the Debian package was becoming a burden, and the lack of synchronization between the distribution made it hard to keep the package up to date.

On the new sysvinit team is the SuSe maintainer Dr. Werner Fink, and my Debian co-maintainer Kel Modderman. About 10 days ago, I made a new upstream tarball with version number 2.87dsf (for Debian, SuSe and Fedora), based on the patches currently in use in these distributions. We Debian maintainers plan to move to this tarball as the new upstream as soon as we find time to do the merge. Since the new tarball was created, we agreed with Werner at SuSe to make a new upstream project at Savannah, and continue development there. The project is registered and currently waiting for approval by the Savannah administrators, and as soon as it is approved, we will import the old versions from svn and continue working on the future release.

It is a bit ironic that this is done now, when some of the involved distributions are moving to upstart as a syvinit replacement.

I spent Monday and tuesday this week in London with a lot of the people involved in the boot system on Debian and Ubuntu, to see if we could find more ways to speed up the boot system. This was an Ubuntu funded developer gathering. It was quite productive. We also discussed the future of boot systems, and ways to handle the increasing number of boot issues introduced by the Linux kernel becoming more and more asynchronous and event base. The Ubuntu approach using udev and upstart might be a good way forward. Time will show.

Anyway, there are a few ways at the moment to speed up the boot process in Debian. All of these should be applied to get a quick boot:

• Use dash as /bin/sh.
• Disable the init.d/hwclock*.sh scripts and make sure the hardware clock is in UTC.
• Install and activate the insserv package to enable dependency based boot sequencing, and enable concurrent booting.

These points are based on the Google summer of code work done by Carlos Villegas.

Support for makefile-style concurrency during boot was uploaded to unstable yesterday. When we tested it, we were able to cut 6 seconds from the boot sequence. It depend on very correct dependency declaration in all init.d scripts, so I expect us to find edge cases where the dependences in some scripts are slightly wrong when we start using this.

On our IRC channel for this effort, #pkg-sysvinit, a new idea was introduced by Raphael Geissert today, one that could affect the startup speed as well. Instead of starting some scripts concurrently from rcS.d/ and another set of scripts from rc2.d/, it would be possible to run a of them in the same process. A quick way to test this would be to enable insserv and run 'mv /etc/rc2.d/S* /etc/rcS.d/; insserv'. Will need to test if that work. :)