Speeding up the boot process (next generation)

This tutorial uses an automated process of generating the "useful" prefetch database.

This is /etc/init.d/my-readahead.
Code:

#!/sbin/runscript
# Copyright 1999-2006 Gentoo Foundation
# Distributed under the terms of the GNU General Public License v2
# $Header: $

depend() {
        before syslog-ng
}

start() {
        ebegin "Starting readahead ROYALE"
        cat /etc/conf.d/* > /dev/null 2>&1 &
        cat /usr/lib/gconv/gconv-modules.cache > /dev/null 2>&1 &
        cat /usr/share/icons/gnome/icon-theme.cache > /dev/null 2>&1 &
        cat /usr/share/icons/hicolor/icon-theme.cache > /dev/null 2>&1 &
        cat `cat /etc/conf.d/sofile-list.load` > /dev/null 2>&1 &
        if [ -f /forcesampler ];then
                /usr/sbin/sample-init-process /etc/conf.d/sofile-list.load &
                rm -f /forcesampler
        fi 
        eend 0
}

stop() {
        eend 0
}

The executable shell script /usr/sbin/sample-init-process:
Code:

 #!/bin/bash
final_db="$1"

# total number of lists to keep in /etc/conf.d/
total_to_keep=4

# total number of samples to make in /forcesampler boot.
# It will take twice this amount of seconds to finish sampling.
total_samples_per_boot=200

# touch empty file if not there
[ ! -f "$final_db" ] && touch "$final_db"

let i=0
while [ "$i" -ne "$total_to_keep" ] ; do
        if [ ! -f "$final_db$i" ] ; then
                break;
        else
                j=$((i+1))
                if [ "$j" = "$total_to_keep" ] ;then
                        j=0
                fi
                if [ -f "$final_db$j" -a "$final_db$i" -nt "$final_db$j" ];then
                        i=$((j))
                        break;
                fi
        fi
        i=$((i+1))
done
if [ "$i" = "$total_to_keep" ] ;then
        i=0
fi

slot="$i"
cp "$final_db" "$final_db$i"

function collect_sample()
{
        lsof | awk '{if ($5=="REG") {if (match($9,".so") || match($9,".conf") \
        || match($9,"/bin/") || match($9,"ttf") || match($9,"libexec")) \
        print $9}}' | sort | uniq >> $1
}

collect_sample /tmp/init_sample1

i=0 
while [ "$i" -ne "$total_samples_per_boot" ] ; do
        collect_sample /tmp/init_sample2
        cat /tmp/init_sample1 /tmp/init_sample2 | sort | uniq > /tmp/init_sample3
        mv /tmp/init_sample3 /tmp/init_sample1
        sleep 1
        i=$((i+1))
done

cat /tmp/init_sample1 "$final_db$slot" | sort | uniq > "$final_db"

rm -f /tmp/init_sample[123]

chmod +x /usr/sbin/sample-init-process
rc-update add my-readahead default
touch /forcesampler

What you have done is asked my-readahead to create a sampler process which samples what files (fonts, binaries, libraries, configs) are in use, every second. It collects 200 samples (will be busy for at least 10 minutes after boot) and keeps updating the database with a list of useful files.

Once you boot after this the first time, you will probably see a slow down because its collecting samples for use on the next and subsequent boots. Use your system as you normally would. Open firefox, gaim, OO or whatever. Once the sampler script has finished ('ps -aef|grep sample-init-process' will tell you if its running), reboot. Now this and all subsequent boots will be much faster.

If it takes too long for the sample-init-process to finish, reduce the total_samples_per_boot variable in the sampler script to 100 and try again by touching /forcesampler.

In future, whenever you feel like updating the prefetch database, just 'touch /forcesampler' and database will be updated on next reboot.

based on http://forums.gentoo.org/viewtopic-t-478491.html, some script modifications done.