docs: rework of chapter "Install with apache"

BTW: normalize installation-nginx.rst
This commit is contained in:
Markus Heiser 2020-04-06 17:59:06 +02:00
parent c81849cb5a
commit eb0d4646d8
7 changed files with 465 additions and 63 deletions

View File

@ -4,19 +4,303 @@
Install with apache
===================
.. sidebar:: public to the internet?
.. _Apache: https://httpd.apache.org/
.. _Apache Debian:
https://cwiki.apache.org/confluence/display/HTTPD/DistrosDefaultLayout#DistrosDefaultLayout-Debian,Ubuntu(Apachehttpd2.x):
.. _README.Debian:
https://salsa.debian.org/apache-team/apache2/raw/master/debian/apache2.README.Debian
.. _Apache Arch Linux:
https://wiki.archlinux.org/index.php/Apache_HTTP_Server
.. _Apache Fedora:
https://docs.fedoraproject.org/en-US/quick-docs/getting-started-with-apache-http-server/index.html
.. _Apache directives:
https://httpd.apache.org/docs/trunk/mod/directives.html
.. _Getting Started:
https://httpd.apache.org/docs/current/en/getting-started.html
.. _Terms Used to Describe Directives:
https://httpd.apache.org/docs/current/en/mod/directive-dict.html
.. _Configuration Files:
https://httpd.apache.org/docs/current/en/configuring.html
.. _ProxyPreserveHost: https://httpd.apache.org/docs/trunk/mod/mod_proxy.html#proxypreservehost
.. _LoadModule:
https://httpd.apache.org/docs/2.4/mod/mod_so.html#loadmodule
.. _DocumentRoot:
https://httpd.apache.org/docs/trunk/mod/core.html#documentroot
.. _Location:
https://httpd.apache.org/docs/trunk/mod/core.html#location
.. _uWSGI Apache support:
https://uwsgi-docs.readthedocs.io/en/latest/Apache.html
.. _apache uwsgi:
https://uwsgi-docs.readthedocs.io/en/latest/Apache.html#mod-proxy-uwsgi
.. _mod_proxy_uwsgi:
https://uwsgi-docs.readthedocs.io/en/latest/Apache.html#mod-proxy-uwsgi
If your searx instance is public, stop here and first install :ref:`filtron
reverse proxy <filtron.sh>` and :ref:`result proxy morty <morty.sh>`, see
:ref:`installation scripts`.
.. sidebar:: further read
- `Apache Arch Linux`_
- `Apache Debian`_ and `README.Debian`_
- `Apache Fedora`_
- `Apache directives`_
.. contents:: Contents
:depth: 2
:local:
:backlinks: entry
Add wsgi mod
============
The apache HTTP server
======================
If Apache_ is not installed, install it now. If apache_ is new to you, the
`Getting Started`_, `Configuration Files`_ and `Terms Used to Describe
Directives`_ documentation gives first orientation. There is also a list of
`Apache directives`_ *to keep in the pocket*.
.. tabs::
.. group-tab:: Ubuntu / debian
.. code:: sh
sudo -H apt-get install apache2
.. group-tab:: Arch Linux
.. code:: sh
sudo -H pacman -S apache
sudo -H systemctl enable httpd
sudo -H systemctl start http
.. group-tab:: Fedora / RHEL
.. code:: sh
sudo -H dnf install httpd
sudo -H systemctl enable httpd
sudo -H systemctl start httpd
Now at http://localhost you should see any kind of *Welcome* or *Test* page.
How this default intro site is configured, depends on the linux distribution
(compare `Apache directives`_).
.. tabs::
.. group-tab:: Ubuntu / debian
.. code:: sh
less /etc/apache2/sites-enabled/000-default.conf
In this file, there is a line setting the `DocumentRoot`_ directive:
.. code:: apache
DocumentRoot /var/www/html
And the *welcome* page is the HTML file at ``/var/www/html/index.html``.
.. group-tab:: Arch Linux
.. code:: sh
less /etc/httpd/conf/httpd.conf
In this file, there is a line setting the `DocumentRoot`_ directive:
.. code:: apache
DocumentRoot "/srv/http"
<Directory "/srv/http">
Options Indexes FollowSymLinks
AllowOverride None
Require all granted
</Directory>
The *welcome* page of Arch Linux is a page showing directory located at
``DocumentRoot``. This is *directory* page is generated by the Module
`mod_autoindex <https://httpd.apache.org/docs/2.4/mod/mod_autoindex.html>`_:
.. code:: apache
LoadModule autoindex_module modules/mod_autoindex.so
...
Include conf/extra/httpd-autoindex.conf
.. group-tab:: Fedora / RHEL
.. code:: sh
less /etc/httpd/conf/httpd.conf
In this file, there is a line setting the ``DocumentRoot`` directive:
.. code:: apache
DocumentRoot "/var/www/html"
...
<Directory "/var/www">
AllowOverride None
# Allow open access:
Require all granted
</Directory>
On fresh installations, the ``/var/www`` is empty and the *default
welcome page* is shown, the configuration is located at::
less /etc/httpd/conf.d/welcome.conf
.. _The Debian Layout:
The Debian Layout
=================
Be aware that the Debian layout is quite different from the standard Apache
configuration. For details look at the README.Debian_
(``/usr/share/doc/apache2/README.Debian.gz``). Some commands you should know on
Debian:
* :man:`apache2ctl`: Apache HTTP server control interface
* :man:`a2enmod`, :man:`a2dismod`: switch on/off modules
* :man:`a2enconf`, :man:`a2disconf`: switch on/off configurations
* :man:`a2ensite`, :man:`a2dissite`: switch on/off sites
.. _apache searx site:
Apache Reverse Proxy
====================
.. sidebar:: public to the internet?
If your searx instance is public, stop here and first install :ref:`filtron
reverse proxy <filtron.sh>` and :ref:`result proxy morty <morty.sh>`, see
:ref:`installation scripts`. If already done, follow setup: *searx via
filtron plus morty*.
To setup a Apache revers proxy you have to enable the *headers* and *proxy*
modules and create a `Location`_ configuration for the searx site. In most
distributions you have to uncomment the lines in the main configuration file,
except in the :ref:`The Debian Layout`.
.. tabs::
.. group-tab:: Ubuntu / debian
In the Apache setup, enable headers and proxy modules:
.. code:: sh
sudo -H a2enmod headers
sudo -H a2enmod proxy
sudo -H a2enmod proxy_http
In :ref:`The Debian Layout` you create a ``searx.conf`` with the
``<Location /searx >`` directive and save this file in the *sites
available* folder at ``/etc/apache2/sites-available``. To enable the
``searx.conf`` use :man:`a2ensite`:
.. code:: sh
sudo -H a2ensite searx.conf
.. group-tab:: Arch Linux
In the ``/etc/httpd/conf/httpd.conf`` file, activate headers and proxy
modules (LoadModule_):
.. code:: apache
LoadModule headers_module modules/mod_headers.so
LoadModule proxy_module modules/mod_proxy.so
LoadModule proxy_http_module modules/mod_proxy_http.so
.. group-tab:: Fedora / RHEL
In the ``/etc/httpd/conf/httpd.conf`` file, activate headers and proxy
modules (LoadModule_):
.. code:: apache
LoadModule headers_module modules/mod_headers.so
LoadModule proxy_module modules/mod_proxy.so
LoadModule proxy_http_module modules/mod_proxy_http.so
.. tabs::
.. group-tab:: searx via filtron plus morty
Use this setup, if your instance is public to the internet, compare
figure: :ref:`architecture <arch public>` and :ref:`installation scripts`.
1. Configure a reverse proxy for :ref:`filtron <filtron.sh>`, listening on
*localhost 4004* (:ref:`filtron route request`):
.. code:: apache
<Location /searx >
# SetEnvIf Request_URI "/searx" dontlog
# CustomLog /dev/null combined env=dontlog
Require all granted
Order deny,allow
Deny from all
#Allow from fd00::/8 192.168.0.0/16 fe80::/10 127.0.0.0/8 ::1
Allow from all
ProxyPreserveHost On
ProxyPass http://127.0.0.1:4004
RequestHeader set X-Script-Name /searx
</Location>
2. Configure reverse proxy for :ref:`morty <searx morty>`, listening on
*localhost 3000* (FYI: ``ProxyPreserveHost On`` is already set, see
above):
.. code:: apache
ProxyPreserveHost On
<Location /morty >
# SetEnvIf Request_URI "/morty" dontlog
# CustomLog /dev/null combined env=dontlog
Require all granted
Order deny,allow
Deny from all
#Allow from fd00::/8 192.168.0.0/16 fe80::/10 127.0.0.0/8 ::1
Allow from all
ProxyPass http://127.0.0.1:3000
RequestHeader set X-Script-Name /morty
</Location>
Note that reverse proxy advised to be used in case of single-user or
low-traffic instances. For a fully result proxification add :ref:`morty's
<searx morty>` **public URL** to your :origin:`searx/settings.yml`:
.. code:: yaml
result_proxy:
# replace example.org with your server's public name
url : https://example.org/morty
server:
image_proxy : True
uWSGI support
=============
Be warned, with this setup, your instance isn't :ref:`protected <searx
filtron>`. Nevertheless it is good enough for intranet usage and it
demonstrates: *how different the uwsgi support is, depending on the
distribution*. To enable :ref:`uWSGI <searx uwsgi>` support you need to install
the apache `apache uwsgi`_ support:
.. tabs::
@ -27,37 +311,130 @@ Add wsgi mod
sudo -H apt-get install libapache2-mod-uwsgi
sudo -H a2enmod uwsgi
Add this configuration in the file ``/etc/apache2/apache2.conf``. To limit
acces to your intranet replace ``Allow from all`` directive and replace
.. group-tab:: Arch Linux
.. code:: sh
sudo -H pacman -S uwsgi
In the ``/etc/httpd/conf/httpd.conf`` file, activate headers and proxy
modules (LoadModule_):
.. code:: apache
LoadModule proxy_module modules/mod_proxy.so
LoadModule proxy_uwsgi_module modules/mod_proxy_uwsgi.so
.. group-tab:: Fedora / RHEL
.. code:: sh
sudo -H dnf install uwsgi
FIXME: enable uwsgi in apache
The next example shows a configuration using the `uWSGI Apache support`_ via
unix sockets. For socket communication, you have to activate ``socket =
/run/uwsgi/app/searx/socket`` and comment out the ``http = 127.0.0.1:8888``
configuration in your :ref:`uwsgi ini file <uwsgi configuration>`.
If not already exists, create a folder for the unix sockets, which can be
used by the searx account:
.. code:: bash
sudo -H mkdir -p /run/uwsgi/app/searx/
sudo -H chown -R searx:searx /run/uwsgi/app/searx/
To limit acces to your intranet replace ``Allow from all`` directive and replace
``192.168.0.0/16`` with your subnet IP/class.
.. _inranet apache site:
.. tabs::
Note that if your instance of searx is not at the root, you should change
``<Location />`` by the location of your instance, like ``<Location /searx>``:
.. group-tab:: Ubuntu / debian
.. code:: apache
Debian uses the (old) `mod_uwsgi
<https://uwsgi-docs.readthedocs.io/en/latest/Apache.html#mod-uwsgi>`_.
# CustomLog /dev/null combined
.. code:: apache
<IfModule mod_uwsgi.c>
<IfModule mod_uwsgi.c>
<Location />
# SetEnvIf Request_URI "/searx" dontlog
# CustomLog /dev/null combined env=dontlog
Options FollowSymLinks Indexes
SetHandler uwsgi-handler
uWSGISocket /run/uwsgi/app/searx/socket
<Location /searx >
Order deny,allow
Deny from all
# Allow from fd00::/8 192.168.0.0/16 fe80::/10 127.0.0.0/8 ::1
Allow from all
Require all granted
</Location>
Options FollowSymLinks Indexes
SetHandler uwsgi-handler
uWSGISocket /run/uwsgi/app/searx/socket
</IfModule>
Order deny,allow
Deny from all
# Allow from fd00::/8 192.168.0.0/16 fe80::/10 127.0.0.0/8 ::1
Allow from all
Enable apache mod_uwsgi and restart apache:
</Location>
</IfModule>
.. group-tab:: Arch Linux
Arch Linux uses the (recommend) `mod_proxy_uwsgi`_.
.. code:: apache
<IfModule proxy_uwsgi_module>
# SetEnvIf Request_URI /searx dontlog
# CustomLog /dev/null combined env=dontlog
<Location /searx>
Require all granted
Order deny,allow
Deny from all
# Allow from fd00::/8 192.168.0.0/16 fe80::/10 127.0.0.0/8 ::1
Allow from all
ProxyPreserveHost On
ProxyPass unix:/run/uwsgi/app/searx/socket|uwsgi://uwsgi-uds-searx/
</Location>
</IfModule>
.. group-tab:: Fedora / RHEL
RHEL uses the (recommend) `mod_proxy_uwsgi`_.
.. code:: apache
<IfModule proxy_uwsgi_module>
# SetEnvIf Request_URI /searx dontlog
# CustomLog /dev/null combined env=dontlog
<Location /searx>
Require all granted
Order deny,allow
Deny from all
# Allow from fd00::/8 192.168.0.0/16 fe80::/10 127.0.0.0/8 ::1
Allow from all
ProxyPreserveHost On
ProxyPass unix:/run/uwsgi/app/searx/socket|uwsgi://uwsgi-uds-searx/
</Location>
</IfModule>
.. _restart apache:
Restart service
===============
.. tabs::
@ -65,31 +442,34 @@ Enable apache mod_uwsgi and restart apache:
.. code:: sh
a2enmod uwsgi
sudo -H systemctl restart apache2
sudo -H service uwsgi restart searx
.. group-tab:: Arch Linux
.. code:: sh
sudo -H systemctl restart httpd
sudo -H systemctl restart uwsgi@searx
.. group-tab:: Fedora / RHEL
.. code:: sh
sudo -H systemctl restart httpd
sudo -H touch /etc/uwsgi.d/searx.ini
disable logs
============
For better privacy you can disable Apache logs. Go back to
``/etc/apache2/apache2.conf`` :ref:`[example] <inranet apache site>` and above
``<Location />`` activate directive:
For better privacy you can disable Apache logs. In the examples above activate
one of the lines and `restart apache`_::
.. code:: apache
CustomLog /dev/null combined
# SetEnvIf Request_URI "/searx" dontlog
# CustomLog /dev/null combined env=dontlog
Restart apache:
.. tabs::
.. group-tab:: Ubuntu / debian
.. code:: sh
sudo -H systemctl restart apache2
.. warning::
You can only disable logs for the whole (virtual) server not for a specific
path.
The ``CustomLog`` directive disable logs for the whole (virtual) server, use it
when the URL of the service does not have a path component (``/searx``) / is
located at root (``/``).

View File

@ -28,10 +28,9 @@ Install with nginx
The nginx HTTP server
=====================
If nginx_ is not installed (uwsgi will not work with the package nginx-light)
If nginx_ is not installed (uwsgi will not work with the package nginx-light),
install it now.
.. tabs::
.. group-tab:: Ubuntu / debian
@ -58,7 +57,8 @@ install it now.
Now at http://localhost you should see a *Welcome to nginx!* page, on Fedora you
see a *Fedora Webserver - Test Page*. The test page comes from the default
`nginx server configuration`_:
`nginx server configuration`_. How this default intro site is configured,
depends on the linux distribution:
.. tabs::
@ -111,7 +111,8 @@ A nginx searx site
If your searx instance is public, stop here and first install :ref:`filtron
reverse proxy <filtron.sh>` and :ref:`result proxy morty <morty.sh>`, see
:ref:`installation scripts`.
:ref:`installation scripts`. If already done, follow setup: *searx via
filtron plus morty*.
Now you have to create a configuration for the searx site. If nginx_ is new to
you, the `nginx beginners guide`_ is a good starting point and the `Getting
@ -143,9 +144,10 @@ Started wiki`_ is always a good resource *to keep in the pocket*.
.. group-tab:: searx via filtron plus morty
Use this setup, if your instance is public to the internet, compare
figure: :ref:`architecture <arch public>`. Configure a reverse proxy for
:ref:`filtron <filtron.sh>`, listening on *localhost 4004* (:ref:`filtron
route request`):
figure: :ref:`architecture <arch public>` and :ref:`installation scripts`.
1. Configure a reverse proxy for :ref:`filtron <filtron.sh>`, listening on
*localhost 4004* (:ref:`filtron route request`):
.. code:: nginx
@ -159,8 +161,8 @@ Started wiki`_ is always a good resource *to keep in the pocket*.
}
Configure reverse proxy for :ref:`morty <searx morty>`, listening on
*localhost 3000*:
2. Configure reverse proxy for :ref:`morty <searx morty>`, listening on
*localhost 3000*:
.. code:: nginx
@ -181,7 +183,10 @@ Started wiki`_ is always a good resource *to keep in the pocket*.
result_proxy:
# replace example.org with your server's public name
url : https://example.org/
url : https://example.org/morty
server:
image_proxy : True
.. group-tab:: proxy or uWSGI
@ -244,7 +249,7 @@ Started wiki`_ is always a good resource *to keep in the pocket*.
mkdir -p /run/uwsgi/app/searx/
sudo -H chown -R searx:searx /run/uwsgi/app/searx/
.. group-tab:: subdirectory URL
.. group-tab:: proxy at subdir URL
Be warned, with these setups, your instance isn't :ref:`protected <searx
filtron>`. The examples are just here to demonstrate how to export the

View File

@ -54,3 +54,9 @@ This installs searx as described in :ref:`installation basic`.
.. code:: bash
$ sudo -H ./utils/morty.sh install all
If all services are running fine, you can add it to your HTTP server:
- :ref:`installation apache`
- :ref:`installation nginx`

View File

@ -634,9 +634,15 @@ if [[ -z "${APACHE_SITES_AVAILABE}" ]]; then
fi
apache_is_installed() {
(command -v apachectl \
&& command -v a2ensite \
&& command -v a2dissite ) &>/dev/null
case $DIST_ID-$DIST_VERS in
ubuntu-*|debian-*)
(command -v apachectl \
&& command -v a2ensite \
&& command -v a2dissite ) &>/dev/null
;;
arch) (command -v httpd) ;;
fedora) (command -v httpd) ;;
esac
}
apache_reload() {

View File

@ -735,6 +735,11 @@ This installs the searx uwsgi app as apache site. If your server is public to
the internet, you should instead use a reverse proxy (filtron) to block
excessively bot queries."
case $DIST_ID-$DIST_VERS in
ubuntu-*|debian-*) : ;;
*) err_msg "sorry distro $DIST_ID $DIST_VERS not yet supported"; exit 42 ;;
esac
! apache_is_installed && err_msg "Apache is not installed."
if ! ask_yn "Do you really want to install apache site for searx-uwsgi?"; then

View File

@ -18,8 +18,8 @@ ProxyPreserveHost On
ProxyPass http://${MORTY_LISTEN}
RequestHeader set X-Script-Name ${PUBLIC_URL_PATH_MORTY}
# In Apache it seems, that setting HTTP_HOST header direct here does have no
# effect. I needed to set 'ProxyPreserveHost On' (see above).
# In Apache it seems, that setting HTTP_HOST header directive here does have
# no effect. I needed to set 'ProxyPreserveHost On' (see above).
# RequestHeader set Host ${PUBLIC_HOST}

View File

@ -23,8 +23,8 @@ ProxyPreserveHost On
ProxyPass http://${FILTRON_LISTEN}
RequestHeader set X-Script-Name ${FILTRON_URL_PATH}
# In Apache it seems, that setting HTTP_HOST header direct here does have no
# effect. I needed to set 'ProxyPreserveHost On' (see above). HTTP_HOST
# In Apache it seems, that setting HTTP_HOST header directive here does have
# no effect. I needed to set 'ProxyPreserveHost On' (see above). HTTP_HOST
# (ProxyPreserveHost On) is needed by searx to render correct *Search URL*
# in the *Link* box and *saved preference*.