Moebooru
Moebooru is a well maintained Danbooru-style image tag system, used by Yande.re, Konachan, and many others. It is written in Ruby, and can be a bit of a challenge to install. Since many other installation guides are insufficient, we've noted down our entire installation and configuration process for Eikonos.
Assumptions
In this guide, we make certain assumptions about your configuration. Your situation may vary.
- Nginx - The webserver used in this guide. Follow the instructions in that page to install Nginx. Then, you can find some sample Nginx configurations for Moebooru in sections below.
* www-data - A group used by all webservers. Make sure to add the nginx user to this group!
moeuser - The unprivileged user that runs this app. After configuration, this will be converted to a nonlogin user./home/moe/booru/- The home folder of the moebooru user, where the application data will be stored from agit clone./home/moe/booru/shared/- The location of the UNIX sockets that puma will use. It is probably safer to put it into/var/run/moebooruinstead, but it is what it is and we have SELinux policies to account for it.
/var/www/moebooru/public/- The public directory will be moved to/var/www/, and symlinked to/home/moe/booru/public/. This way, Nginx will not have control over the entire application directory.- Make sure to restore this symlink every time an upgrade is run with
git pull.
- Make sure to restore this symlink every time an upgrade is run with
Create Moebooru User
Create a specific non-login daemon user just for moebooru (Though it will have bash shell for setup purposes temporarily).
sudo adduser moe # create whatever password
sudo passwd -d moe # delete password to prevent normal login
sudo -i -u moe # become the moe user
git clone https://github.com/moebooru/moebooru.git /home/moe/booru
chown -R moe:moe /home/moe/booru
As root, we also need to move all publicly served assets to /var/www/.
sudo mkdir -p /var/www/moebooru
sudo mv /home/moe/booru/public /var/www/moebooru/
sudo ln -s /var/www/moebooru/public /home/moe/booru/public
sudo chown moe:moe /home/moe/booru/public
sudo chmod 755 /var/www/moebooru/public
sudo chown root:www-data -R /var/www/moebooru
Setup Postgresql
The examples below use PostgreSQL 9.6, but as of 2017-01-28, Moebooru just needs greater than 9.4.
Debian: (Alternatively, you can get it from jessie-backports and newer)
/etc/apt/sources.list.d/pgdg.list
deb http://apt.postgresql.org/pub/repos/apt/ YOUR_DEBIAN_VERSION_HERE-pgdg main
wget --quiet -O - https://www.postgresql.org/media/keys/ACCC4CF8.asc
# yum install http://yum.postgresql.org/9.6/redhat/rhel-7-x86_64/pgdg-redhat95-9.5-2.noarch.rpm
# yum install postgresql96 postgresql96-devel postgresql96-server libpqxx libpqxx-devel
libpqxx is required for libpq-ruby to build.
Now log in and create the moebooru database and user: (make sure to create a very good password!)
$ sudo su # becoming the postgres user doesn't really work with sudo for some reason
# su - postgres
$ psql
postgres # create user moebooru_user with password 'your_password' createdb;
Finally, edit pg_hba.conf from ident to md5 (except for the UNIX socket line) to allow users to log in using a password (required by moebooru's config), it should look like the following:
RHEL/CentOS: /var/lib/pgsql/9.6/data/pg_hba.conf Debian: /etc/postgresql/9.6/main/pg_hba.conf
# "local" is for Unix domain socket connections only local all all md5 # IPv4 local connections: host all all 127.0.0.1/32 md5 # IPv6 local connections: host all all ::1/128 md5 # Allow replication connections from localhost, by a user with the # replication privilege. #local replication postgres peer #host replication postgres 127.0.0.1/32 trust #host replication postgres ::1/128 trust
Once you have this file edited as seen above, restart postgresql.
sudo systemctl restart postgresql-9.6
Install Nodejs
NodeJS is necessary for the frontend. You should obtain the latest version, 7.x: by following these guides for your distro.
Setup Ruby
For most small sites, Ruby 2.3 and higher with the Unicorn server is sufficient.
A site with heavier traffic may find it helpful to use Rubinius with the Puma server make better use of concurrent multithreading, but this may require compilation.
Method 1: Set up Normal Ruby 2.3 or newer
Most distros have an outdated version of Ruby, so you must set up repositories from the official developers:
- Debian: Brightbox PPA - Ruby 2.3
- RHEL/CentOS: RHSCL - Ruby 2.3
Method 2: Install Rubinius with RVM
For RVM, just install any ol' ruby 2.x, we won't be using it after the compilation stage. Also install all the build dependencies.
Dependencies (Debian/Ubuntu)
sudo apt-get install ruby ruby-dev
sudo apt-get install build-essential openssl libreadline6 libreadline6-dev \
curl git-core zlib1g zlib1g-dev libssl-dev libyaml-dev libsqlite3-dev sqlite3 \
libxml2-dev libxslt-dev autoconf libc6-dev ncurses-dev automake libtool bison \
subversion pkg-config libgdbm-dev ncurses-dev libffi-dev
An LLVM version greater than 3.6+ is also required. You can get this in Debian Jessie from backports, or in Ubuntu from normal repositories.
# if on debian jessie, set up jessie-backports: https://backports.debian.org/Instructions/
sudo apt-get install llvm-3.8 clang-3.8 libclang-3.8-dev llvm-3.8-runtime llvm-3.8-dev llvm-3.8-tools libedit2 libedit-dev
Click Expand to display the update-alternatives commands, which activate 3.8 as the default clang/llvm version.
sudo update-alternatives --install /usr/bin/clang++ clang++ /usr/bin/clang++-3.8 100
sudo update-alternatives --install /usr/bin/clang clang /usr/bin/clang-3.8 100
sudo update-alternatives --install \
/usr/bin/llvm-config llvm-config /usr/bin/llvm-config-3.8 200 \
--slave /usr/bin/llvm-ar llvm-ar /usr/bin/llvm-ar-3.8 \
--slave /usr/bin/llvm-as llvm-as /usr/bin/llvm-as-3.8 \
--slave /usr/bin/llvm-bcanalyzer llvm-bcanalyzer /usr/bin/llvm-bcanalyzer-3.8 \
--slave /usr/bin/llvm-cov llvm-cov /usr/bin/llvm-cov-3.8 \
--slave /usr/bin/llvm-diff llvm-diff /usr/bin/llvm-diff-3.8 \
--slave /usr/bin/llvm-dis llvm-dis /usr/bin/llvm-dis-3.8 \
--slave /usr/bin/llvm-dwarfdump llvm-dwarfdump /usr/bin/llvm-dwarfdump-3.8 \
--slave /usr/bin/llvm-extract llvm-extract /usr/bin/llvm-extract-3.8 \
--slave /usr/bin/llvm-link llvm-link /usr/bin/llvm-link-3.8 \
--slave /usr/bin/llvm-mc llvm-mc /usr/bin/llvm-mc-3.8 \
--slave /usr/bin/llvm-mcmarkup llvm-mcmarkup /usr/bin/llvm-mcmarkup-3.8 \
--slave /usr/bin/llvm-nm llvm-nm /usr/bin/llvm-nm-3.8 \
--slave /usr/bin/llvm-objdump llvm-objdump /usr/bin/llvm-objdump-3.8 \
--slave /usr/bin/llvm-ranlib llvm-ranlib /usr/bin/llvm-ranlib-3.8 \
--slave /usr/bin/llvm-readobj llvm-readobj /usr/bin/llvm-readobj-3.8 \
--slave /usr/bin/llvm-rtdyld llvm-rtdyld /usr/bin/llvm-rtdyld-3.8 \
--slave /usr/bin/llvm-size llvm-size /usr/bin/llvm-size-3.8 \
--slave /usr/bin/llvm-stress llvm-stress /usr/bin/llvm-stress-3.8 \
--slave /usr/bin/llvm-symbolizer llvm-symbolizer /usr/bin/llvm-symbolizer-3.8 \
--slave /usr/bin/llvm-tblgen llvm-tblgen /usr/bin/llvm-tblgen-3.8
Dependencies (RHEL/CentOS)
sudo yum install ruby
sudo yum install -y patch autoconf patch automake libtool bison sqlite-devel
One special build dependency is llvm-3.6+. Unfortunately, CentOS 7's EPEL repo only has 3.4, so we will need to get the latest build of LLVM from Fedora COPR.
/etc/yum.repos.d/daveisfera-llvm_3.7-epel-7.repo
[daveisfera-llvm_3.7] name=Copr repo for llvm_3.7 owned by daveisfera baseurl=https://copr-be.cloud.fedoraproject.org/results/daveisfera/llvm_3.7/epel-7-$basearch/ type=rpm-md skip_if_unavailable=True gpgcheck=1 gpgkey=https://copr-be.cloud.fedoraproject.org/results/daveisfera/llvm_3.7/pubkey.gpg repo_gpgcheck=0 enabled=1 enabled_metadata=1
Then just install llvm as normal:
sudo yum install clang llvm llvm-devel llvm-static libedit libedit-devel
Using RVM
Follow the RVM installation instructions, as replicated below:
sudo -i -u moe
gpg --keyserver hkp://keys.gnupg.net --recv-keys 409B6B1796C275462A1703113804BB82D39DC0E3
curl -sSL https://get.rvm.io | bash -s stable
Then, log out and log back in to reload bash and make the RVM command work:
exit
sudo -i -u moe
Refresh the rvm repos. And since we have already installed dependencies in the sections above, disable the built-in dependency resolver:
rvm get head
rvm autolibs disable
Install rubinius. Compilation will take a while.
rvm install rbx
After installation, choose Rubinius (rbx) as the default ruby version to use.
rvm list
This will list out various ruby versions.
rvm alias create default rbx
Finally, log out and log back in, and check your current ruby version to ensure that Rubinius is default.
ruby -v
Now that your ruby type is set up, install bundler.
gem install bundler
Mandatory Access Control
AppArmor (Debian/Ubuntu)
anyone have apparmor profiles?
We may have to create our own through learning mode: http://www.howtogeek.com/118328/how-to-create-apparmor-profiles-to-lock-down-programs-on-ubuntu/
SELinux Permissions (RHEL/CentOS)
If using SELinux (which we highly recommend), you will need the following policies, assuming that everything is installed to /home/moe/booru/:
needed for proxy pass
sudo chcon -Rt httpd_sys_content_t /var/www/ # allow nginx to access folders sudo setsebool httpd_can_network_connect 1 -P # allows reverse proxy sudo setsebool -P httpd_can_network_memcache 1 # allows memcache
needed to serve from moebooru user's directory, but only shared/ and public folders
setsebool -P httpd_enable_homedirs 1 sudo semanage fcontext -a -t httpd_sys_content_t '/home/moe/booru/shared(/.*)?' sudo restorecon -R -v /home/moe/booru/shared sudo semanage fcontext -a -t httpd_sys_content_t '/var/www/moebooru/public(/.*)?' sudo restorecon -R -v /var/www/moebooru/public
https://www.pckr.co.uk/selinux-nginx-and-reverse-proxying-2/
The last step is to run the final allows.
sudo grep nginx /var/log/audit/audit.log | audit2allow -M nginx > nginx.te
Open up the nginx.te file and see that it is correct (such that no suspicious rule allows are inside). Then run:
sudo grep nginx /var/log/audit/audit.log | audit2allow -M nginx sudo semodule -i nginx.pp
http://axilleas.me/en/blog/2013/selinux-policy-for-nginx-and-gitlab-unix-socket-in-fedora-19/
Setup Moebooru
Now, conduct the setup and get dependencies.
Dependencies (Debian/Ubuntu)
sudo apt-get install build-essential libxml2-dev libxslt1-dev libpq-dev git jhead libgd2-noxpm-dev imagemagick
- build-essential - a meta-package that installs various tools and libraries necessary for program (mostly C++ and Perl) compilation and interpretation.
- libxml2 - library that is required for XML support. It is necessary for moebooru’s external API to work.
- libxslt1 - extension of libxml that allows better xml parsing and converting.
- libpq-dev - library that is necessary to build postgres connection module for ruby.
- git - version control system - we will use it for moebooru installation and updating.
- jhead - image processing tool that extracts EXIF information from image files.
- libgd2-noxpm - an image processing library.
- libgs2-noxpm-dev - an extension for compiling GD2 support for applications.
- imagemagick - collection of tools for image processing
Dependencies (RHEL/CentOS)
sudo yum install gcc gcc-c++ ImageMagick jhead libxslt-devel git libyaml-devel openssl-devel pcre-devel readline-devel
Configure Moebooru
All commands in this section must be run as the moe user:
sudo -i -u moe
First, do a bundle config for pg since we're using a specific postgresql version:
bundle config build.pg --with-pg-config=/usr/pgsql-9.6/bin/pg_config
Generate your secret key, which is used for salts and such.
bundle exec rake secret
Install the ruby packages for the moebooru user only (under the directory ./vendor/bundle):
bundle install --path vendor/bundle
Obtain config/database.yml and config/local_config.rb from the .example files, and configure them accordingly. Then set chmod 700 so only the moebooru user can read the database password.
cp config/database.yml.example config/database.yml
cp config/local_config.rb.example config/local_config.rb
chmod 700 config/database.yml
chmod 700 config/local_config.rb
Edit config/database.yml using your favorite editor (such as nano), and replace imouto with the password of your database user:
config/database.yml
login: &login adapter: postgresql username: moe password: imouto host: 127.0.0.1 development: database: moebooru_dev <<: *login test: database: moebooru_test <<: *login production: database: moebooru <<: *login
chmod 600 /home/moe/booru/config/database.yml
Initialize database with this command (there will be some errors reported which is expected)
bundle exec rake db:reset
Then, migrate the database tables.
bundle exec rake db:migrate
Now, you need to provide the correct permissions to the public folder:
chmod 755 /home/moe/booru/public
Start the server (bundle exec unicorn or bundle exec puma if using JRuby/Rubinius)
Customize Header Image and Branding
By default, Moebooru comes with the Yande.re header image and branding, as the site developed the moebooru engine. You should definitely consider removing the original branding unless your site is private.
app/assets/images
public/favicon.ico
Enable Memcached
Memcached is a high performance caching solution and is needed to have Moebooru enumerate posts and create the /posts pagination bar. Follow these instructions to install Memcached. You need at least 2GB free RAM to provide.
Installation (Debian/Ubuntu)
sudo apt-get install memcached
https://kyup.com/tutorials/install-use-memcache/
Installation (RHEL/CentOS)
sudo yum install memcached
http://www.liquidweb.com/kb/how-to-install-memcached-on-centos-7/
Configuration
Edit /etc/sysconfig/memcached and set CACHESIZE=2048 (2GB RAM) if possible.
Then set memcached to start at every boot:
sudo systemctl restart memcached
Config-based Activation
Add these options to the following file. When you start moebooru again, memcached will be active.
config/local_config.rb
# The server and port where the memcache client can be accessed. Only relevant if you enable caching. CONFIG["memcache_servers"] = ["localhost:11211"] # This enables various caching mechanisms. You must have memcache (and the memcache-client ruby gem) installed in order for caching to work. CONFIG["enable_caching"] = true
ENV-based Memcached Activation
Activate it by appending a bash variable to the puma command: MB_MEMCACHE_SERVERS="<url-to-memcached>" . Here are some examples.
TCP: MB_MEMCACHED_SERVERS='127.0.0.1:11211' RAILS_ENV=production bundle exec puma -e production
UNIX Socket: MB_MEMCACHED_SERVERS='127.0.0.1:11211' RAILS_ENV=production bundle exec puma -C shared/puma.rb
SELinux Permissions
You will probably need to allow it through SELinux:
https://major.io/2011/09/07/getting-apache-php-and-memcached-working-with-selinux/
Production Mode
By default, Moebooru runs in development mode, which can be slow (since it's designed to allow debug and automatic recompilation). Here are the steps to set up Production mode.
Preparation
First, you need to create the database, and pregenerate the javascript/css (do this every time you update):
RAILS_ENV=production bundle exec rake db:reset RAILS_ENV=production bundle exec rake assets:precompile
Then, you need to provide the correct permissions to the public folder:
chmod 755 /var/www/moebooru/public
Serve static files with Nginx
Create an Nginx config under /etc/nginx/conf.d/. Make sure to change the server_name.
Note: If you are using a different port for puma, (by adding
-p 3000to the serving command), also change the port accordingly below.
server {
listen 80;
server_name booru.eikonos.org;
# directory of static assets, first generate with the command:
# RAILS_ENV=production bundle exec rake assets:precompile
root /var/www/moebooru/public;
try_files $uri/index.html $uri @app;
location @app {
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
# Fix the "It appears that your reverse proxy set up is broken" error.
proxy_pass http://127.0.0.1:9292;
proxy_read_timeout 90;
proxy_redirect http://127.0.0.1:9292 http://$server_name;
}
error_page 500 502 503 504 /500.html;
client_max_body_size 4G;
keepalive_timeout 10;
}
Run the Server
Finally, to run the server (default is port 9292), run one of the following commands:
Ruby:
bundle exec unicorn -p 9292 -E production
Rubinius:
RAILS_ENV=production bundle exec puma -e production
Serve with UNIX Socks in Production Mode
for even more effectiveness, use a UNIX sock: https://www.digitalocean.com/community/tutorials/how-to-deploy-a-rails-app-with-puma-and-nginx-on-ubuntu-14-04
Figure out the amount of CPU cores you have:
grep -c processor /proc/cpuinfo
Create the following folders in your application directory:
mkdir -p shared/pids shared/sockets shared/log
Place the following into <app_dir>/shared/puma.rb:
shared/puma.rb
# Change to match your CPU core count
workers 8
# Min and Max threads per worker
threads 1, 6
app_dir = File.expand_path("../..", __FILE__)
shared_dir = "#{app_dir}/shared"
# Default to production
rails_env = ENV['RAILS_ENV'] || "production"
environment rails_env
# Set up socket location
bind "unix://#{shared_dir}/sockets/puma.sock"
# Logging
stdout_redirect "#{shared_dir}/log/puma.stdout.log", "#{shared_dir}/log/puma.stderr.log", true
# Set master PID and state locations
pidfile "#{shared_dir}/pids/puma.pid"
state_path "#{shared_dir}/pids/puma.state"
activate_control_app
on_worker_boot do
require "active_record"
ActiveRecord::Base.connection.disconnect! rescue ActiveRecord::ConnectionNotEstablished
ActiveRecord::Base.establish_connection(YAML.load_file("#{app_dir}/config/database.yml")[rails_env])
end
Change the Nginx server config to the following:
/etc/nginx/conf.d/booru.eikonos.org.conf
upstream app {
# Path to Puma SOCK file, as defined previously
server unix:/home/moe/booru/shared/sockets/puma.sock fail_timeout=0;
}
server {
listen 80;
server_name booru.eikonos.org;
# directory of static assets, first generate with the command:
# RAILS_ENV=production bundle exec rake assets:precompile
root /var/www/moebooru/public;
try_files $uri/index.html $uri @app;
location @app {
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
# Fix the "It appears that your reverse proxy set up is broken" error.
proxy_pass http://app;
proxy_read_timeout 90;
proxy_redirect http://app http://$server_name;
}
error_page 500 502 503 504 /500.html;
client_max_body_size 4G;
keepalive_timeout 10;
}
Finally, to run the server, use the following:
bundle exec puma -C shared/puma.rb
Systemd Service
https://github.com/puma/puma/blob/master/docs/systemd.md
Save this to /usr/systemd/system/moebooru.service. There are two versions, one for TCP and one for unix socket. Change the WorkingDirectory accordingly.
TCP
/usr/systemd/system/moebooru.service
[Unit] Description=Moebooru's Puma HTTP Server Requires=postgresql-9.6.service Wants=postgresql-9.6.service memcached.service After=network.target postgresql-9.6.service # Uncomment for socket activation (see below) # Requires=puma.socket [Service] # Foreground process (do not use --daemon in ExecStart or config.rb) Type=simple # Preferably configure a non-privileged user User=moe # Specify the path to your puma application root WorkingDirectory=/home/moe/booru # Helpful for debugging socket activation, etc. # Environment=PUMA_DEBUG=1 # The command to start Puma # Here we are using a binstub generated via: # `bundle binstubs puma --path ./sbin` # in the WorkingDirectory (replace <WD> below) # You can alternatively use `bundle exec --keep-file-descriptors puma` # ExecStart=<WD>/sbin/puma -b tcp://0.0.0.0:9292 -b ssl://0.0.0.0:9293?key=key.pem&cert=cert.pem # Alternatively with a config file (in WorkingDirectory) and # comparable `bind` directives ExecStart=/bin/bash -c 'RAILS_ENV=production /usr/bin/bundle exec puma -e production' Restart=always [Install] WantedBy=multi-user.target
UNIX Socket
Two Systemd services are needed: one for the socket and one for the application.
/usr/systemd/system/moebooru.service
[Unit] Description=Puma HTTP Server Requires=postgresql-9.6.service Wants=postgresql-9.6.service memcached.service After=network.target postgresql-9.6.service # Uncomment for socket activation (see below) # Requires=puma.socket [Service] # Foreground process (do not use --daemon in ExecStart or config.rb) Type=simple # Preferably configure a non-privileged user # User= # Specify the path to your puma application root # WorkingDirectory= # Helpful for debugging socket activation, etc. # Environment=PUMA_DEBUG=1 # The command to start Puma # Here we are using a binstub generated via: # `bundle binstubs puma --path ./sbin` # in the WorkingDirectory (replace <WD> below) # You can alternatively use `bundle exec --keep-file-descriptors puma` # ExecStart=<WD>/sbin/puma -b tcp://0.0.0.0:9292 -b ssl://0.0.0.0:9293?key=key.pem&cert=cert.pem # Alternatively with a config file (in WorkingDirectory) and # comparable `bind` directives # ExecStart=<WD>/sbin/puma -C config.rb Restart=always [Install] WantedBy=multi-user.target
Grab some code from here?
https://github.com/puma/puma/issues/976
SSL Certificates
While this is beyond the scope of this guide, you should strongly consider using SSL certificates, which are now free with Let's Encrypt.
https://www.digitalocean.com/community/tutorials/how-to-secure-nginx-with-let-s-encrypt-on-centos-7
Full Backup
PostgreSQL Database (run as moebooru user!):
pg_dump -U moebooru -W moebooru > eikonos.org-20170128.pg.sql # will ask for password
Images only:
tar -cvzf eikonos.org-20170128_images.tar.gz /var/www/moebooru/public/data/image/
Code only, no binaries:
tar --exclude='/home/moe/booru/tmp' --exclude='/home/moe/booru/vendor' -cvjf eikonos.org-20170128_var_www_booru-eikonos-org.tar.gz /home/moe/booru
Sources
- Get configuration tips from here: http://wiki.douglasqsantos.com.br/doku.php/deploying_a_rails_app_on_debian_jessie_with_capistrano_nginx_and_puma_en
