PHP-FPM

If you are going to use PHP 5.3.3 or higher you don’t need to patch your php sources by PHP-FPM patch anymore.
Install all required debian packages:

 apt-get install libxml2-dev \
libxml2 libssl-dev libevent-dev libbz2-dev libcurl4-openssl-dev \
libjpeg62-dev libpng12-dev libxpm-dev libfreetype6-dev libt1-dev \
libmcrypt-dev libmysql++-dev libmysqld-dev libmysqlclient-dev \
libxslt1-dev autoconf 

Then download and unzip your php sources and go inside the dir and run buildconf:

./buildconf --force

if no errors occur run configure:

./configure \
--prefix=/opt/php5 \
--with-config-file-path=/opt/php5/etc \
--with-curl \
--with-pear \
--with-gd \
--with-jpeg-dir \
--with-png-dir \
--with-zlib \
--with-xpm-dir \
--with-freetype-dir \
--with-t1lib \
--with-mcrypt \
--with-mhash \
--with-mysql \
--with-mysqli \
--with-pdo-mysql \
--with-openssl \
--with-xmlrpc \
--with-xsl \
--with-bz2 \
--with-gettext \
--with-fpm-user=www-data \
--with-fpm-group=www-data \
--enable-fpm \
--enable-exif \
--enable-wddx \
--enable-zip \
--enable-bcmath \
--enable-calendar \
--enable-ftp \
--enable-mbstring \
--enable-soap \
--enable-sockets \
--enable-sysvmsg \
--enable-sysvsem \
--enable-sysvshm \
--with-libevent-dir=/usr/lib/

After finish without errors proceed and compile and install it:

make && make install

The place where you can find all php-fpm related files and default prefix is /opt/php5. Now customize your configuration at /opt/php5/etc/php-fpm.conf

[global]
pid = /opt/php5/var/run/php-fpm.pid
error_log = /opt/php5/var/log/php-fpm.log
log_level = warning
daemonize = yes # set yes to make php work as a daemon

[www]
listen = 127.0.0.1:9000 # the ip and port for php to listen; the port which nginx are going to work with
user = www-data # set user of new php instances(children)
group = www-data
pm = dynamic
pm.max_children = 50 # the maximum concurrent user's request which can be served by php
pm.min_spare_servers = 5 # the minimum idle processes
pm.max_spare_servers = 35

More details about how to children and servers work you can get here pm.max_children, pm.max_spare_servers, pm.min_spare_servers. Its works for apache just in the same way as for php-fpm.

The next step is making init scripts… Okay I’ve done it for you. Just enjoy:

#!/bin/bash

### BEGIN INIT INFO
# Provides:          php-fpm
# Required-Start:    $local_fs $remote_fs $network $syslog
# Required-Stop:     $local_fs $remote_fs $network $syslog
# Default-Start:     2 3 4 5
# Default-Stop:      0 1 6
# Short-Description: starts php-fpm daemon
# Description:       starts php-fpm daemon
### END INIT INFO

function start {
    $DAEMON -c $CONFIG &> /dev/null
}

function stop {
#    killall php-fpm &> /dev/null
    kill `cat /opt/php5/var/run/php-fpm.pid 2>/dev/null` &> /dev/null
}

. /lib/lsb/init-functions

CONFIG=/opt/php5/etc/php-fpm.conf
DAEMON=/opt/php5/sbin/php-fpm
DESC="php-fpm daemon"
PID=/opt/php5/var/run/php-fpm.pid

test -x $DAEMON || exit 0
test -d $CONFIG_DIR || exit 0

kill -0 `cat /opt/php5/var/run/php-fpm.pid 2>/dev/null` &> /dev/null
running=$?
case $1 in
start)
    if [ 0 -eq $running ]; then
        echo "Process php-fpm is already running... Nothing to do..."
        exit
    fi
    start && echo "Process php-fpm was successfully started ... "
;;
stop)
    if [ 0 -ne $running ]; then
        echo "Process php-fpm is not running..."
        exit
    fi
    stop && echo "Process php-fpm was successfully stoped ... "
;;
restart)
    stop && start;
    echo "Process php-fpm was successfully restarted ... "
;;
status)
    if [ 0 -eq $running ]; then
        echo "Process php-fpm is running..."
    else
        echo "Process php-fpm is not running..."
    fi
;;
*)
    echo "$0 [start|stop|restart|status]"
;;
esac

Then make the stuff work after reboot add it to runlevels:

update-rc.d php-fpm defaults

NGINX

With nginx installation there is easier way. Simply install it with apt-get package manager:

apt-get install nginx 

I’m not going to describe all configuration just mention about php part. To make nginx work with php let it know about. Go to server section and add:

location ~ \.php$ {
    fastcgi_pass   127.0.0.1:9000;
    fastcgi_index  index.php;
    fastcgi_param  SCRIPT_FILENAME  /var/www/nginx$fastcgi_script_name;
    include        fastcgi_params;
    }

The fastcgi_pass should be the similar to section listen in php-fpm.conf.
Directive fastcgi_param you need to specify path to your dir where php files are located(i.e. for me its /var/www/nginx)

VERIFICATION

To verify that everything work well. Use an old method with phpinfo function. Create php file in you htroot direcory with next content:

<?php
phpinfo();
?>

Then give the name like info.php or whatever(but should with .php extension anyway). And try to open in your browser. If works it would be looks like that.

PROFIT…

To make local https server visible to outside world. You can make PAT(Port Address Translation) but what if 443 port is already binded? Than you probably want to make a reverse proxy like I did. So to configure apache’s virtual domain to work with ssl http do several easy steps.

First, you need to load mod_proxy’s modules into apache and make them available after the restart(for debian there is kind of exotic way to do that):

cd /etc/apache2/mods-enabled
ln -s ../mods-available/proxy.conf proxy.conf
ln -s ../mods-available/proxy.load proxy.load
ln -s ../mods-available/proxy_connect.load proxy_connect.load
ln -s ../mods-available/proxy_http.load proxy_http.load

The idea that contents of each file put in mods-enabled becomes a part of main configuration after restart or configuration re-read.

For example, if file contains line like that

LoadModule proxy_module /usr/lib/apache2/modules/mod_proxy.so

It is the same if we put this line directly in apache2.conf or httpd.conf of whatever you have.

Hope it is pretty clear. Than let’s move on and create new virtual host configuration file:

cd /etc/apache2/sites-available
touch ssl-proxy

Here the example how it’s should be looks like:

<VirtualHost *:443>

  ServerName mail.center.ua
  ProxyRequests Off
  <Proxy *>
    Order deny,allow
    Allow from all
  </Proxy>

  SSLProxyEngine On
  ProxyPass / https://192.168.1.3/
  ProxyPassReverse / https://192.168.1.3/

  # logging section
  LogLevel info
  ErrorLog /var/log/apache2/ssl-proxy.log
  CustomLog /var/log/apache2/ssl-proxy.log combined

  SSLEngine on
  SSLProtocol all
  SSLCertificateFile /etc/ssl/apache/server.crt
  SSLCertificateKeyFile /etc/ssl/apache/server.key
</VirtualHost>

About SSLCertificateFile and SSLCertificateKeyFile you need to generate keys. Use google to figure how to do it.

The mechanism of virtual host is similar to mods mechanism. So create new link in sites-enabled:

ln -s ../sites-available/ssl-proxy ssl-proxy

And restart apache:

/etc/init.d/apache2 restart

Good look!

Get all required software over apt-get:

# apt-get install kernel-package ncurses-dev fakeroot wget bzip2 module-init-tools initrd-tools procps

Next, get latest linux kernel from kernel.org :

# cd /usr/src
# wget http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.30.tar.bz2

you can also done it over apt-get as well:

# apt-get install linux-source

Extract it:

# tar -xjvf linux-2.6.30.tar.bz2

and make symbolic link pointed out your sources and go there:

# ln -s linux-2.6.30.tar.bz2 linux
# cd linux

Configure your new kernel with

make menuconfig

or

make oldconfig

or

# make xconfig

and so on.

After you will finished. Run :

# make-kpkg clean
# fakeroot make-kpkg --revision=kernel_name-version kernel_image

Where is kernel_name-version – kernel name and version(e.g.: custom-1.0).

If compilation done without errors. Go to the /usr/src and install new kernel over

dpkg

tool:

# cd /usr/src
# dpkg -i kernel-image-2.6.30.2_custom-1.0_i386.deb

Create ramdisk from new kernel :

# cd /boot/
# mkinitramfs -o /boot/initrd.img-2.6.30 2.6.30

Last one step is loader configuration. I will describe how to do this for two most popular(grub and lilo) loaders. If you have lilo loader just add new section into the /etc/lilo.conf and edit directive default. Like this:

default=MYKERN
image=/vmlinuz
        label=MYKERN
        read-only
        initrd=/boot/initrd.img-2.6.30

Now, install new loader at boot sector:

# lilo -c /etc/lilo.conf

In case of GRUB loader. Put next section into the /boot/grub/menu.lst file:

title           My Debian GNU/Linux, kernel 2.6.30
root            (hd0,0)
kernel          /boot/vmlinuz-2.6.30 root=/dev/sda1 ro
initrd          /boot/initrd.img-2.6.30

For GRUB loader no installs is require.

After loader section done well you have to restart you machine:

# reboot

Check you current kernel version with uname if during your booting time no problems were occurred:

# uname -r
2.6.30

If it’s matched with you installed kernel version then you done everything right.

NOTICE: In my case distro is Debian.

My current kernel version is :

# uname -r
2.6.30-1-686

Let’s have a look about virtualbox module :

# apt-cache search virtualbox | grep [you_kernel_version]
# 

Nothing was found.

First of all we have to get kernel sources:

# apt-get install linux-source-[you_kernel_version]

Go to /usr/src and delete symbolic link linux if it exist:

# cd /usr/src
# rm linux 

extract files from archive and make new one linx:

# tar -xjvf linux-source-2.6.30.tar.bz2
# ln -s linux-source-2.6.30 linux

Now go to sources and set current configuration:

# cd linux
# make oldconfig && make prepare

After finished, run :

# /etc/init.d/vboxdrv setup
Stopping VirtualBox kernel module:done..
Recompiling VirtualBox kernel module:done..
Starting VirtualBox kernel module:done..

Check is everything done right:

# lsmod | grep -i vbox

It’s okay if you get something like this:

vboxdrv                63936  0

Now you are ready to run VirtualBox.