FoolFuuka/Install/Debian

From Bibliotheca Anonoma
Revision as of 00:53, 19 March 2017 by Antonizoon (talk | contribs) (→‎Extra configuration)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

This guide was written by ATC of desuarchive, and is designed for Debian/Ubuntu Linux servers.

Prepare things

apt-get update
apt-get upgrade

Add the keys for HipHop Virtual Machine

sudo apt-key adv --recv-keys --keyserver hkp://keyserver.ubuntu.com:80 0x5a16e7281be7a449
sudo add-apt-repository "deb http://dl.hhvm.com/ubuntu $(lsb_release -sc) main"

Add repo for MariaDB 10.1

sudo apt-key adv --recv-keys --keyserver hkp://keyserver.ubuntu.com:80 0xcbcb082a1bb943db
sudo add-apt-repository 'deb [arch=amd64,i386] http://nyc2.mirrors.digitalocean.com/mariadb/repo/10.1/ubuntu trusty main'

Add repo for sphinxsearch

Note: Weird shit happens here
sudo add-apt-repository ppa:builds/sphinxsearch-stable

Then, Comment out repo in /etc/apt/sources.list.d/* and add to end:

deb http://ppa.launchpad.net/builds/sphinxsearch-stable/ubuntu precise main 
deb-src http://ppa.launchpad.net/builds/sphinxsearch-stable/ubuntu precise main 
sudo apt-get update

Install all the needed packages.

Nginx, HHVM, PHP5 Cli, PHP5 Curl, MariaDB, OpenJDK w/ compilier, Maven, Sphinxsearch

sudo apt-get install git nginx hhvm php5-fpm php5-cli php5-curl php5-mysqlnd mariadb-server libmariadbclient-dev mariadb-client imagemagick openjdk-7-jre openjdk-7-jdk openjdk-7-jre-lib maven sphinxsearch

Install composer for downloading FoolFuuka dependencies.

curl -sS https://getcomposer.org/installer | php

Move composer to bin

sudo mv composer.phar /usr/local/bin/composer

Install hhvm as a service so it launches at boot.

sudo update-rc.d hhvm defaults

Install hhvm with nginx

sudo /usr/share/hhvm/install_fastcgi.sh

Edit HHVM config to use a unix socket.

sudo nano /etc/hhvm/server.ini

Then remove the line starting with hhvm.server.port, and in its place add the following one:

hhvm.server.file_socket=/var/run/hhvm/hhvm.sock

Set required php.ini vars.

sudo nano /etc/php5/fpm/php.ini
date.timezone = UTC
upload_max_filesize = 32M
post_max_size = 32M

Restart HHVM

sudo service hhvm restart

Configure NGINX to use hhvm's new unix socket.

sudo nano /etc/nginx/hhvm.conf

In this file make sure that the fastcgi_pass directive points to the HHVM socket and looks like this:

  fastcgi_pass unix:/var/run/hhvm/hhvm.sock;

Download FoolFuuka in home folder and install dependencies.

cd ~
composer create-project foolz/foolfuuka foolfuuka --prefer-source
cd foolfuuka
composer dump-autoload --optimize
composer install 

Install Foolfuuka Plugins

You may need to find the repo on their github if their site doesn't work: https://github.com/FoolCode

cd ~/foolfuuka/app/foolz/foolframe/plugins
mkdir foolz
cd foolz
git clone https://github.com/FoolCode/foolframe-plugin-articles.git


cd ~/foolfuuka/app/foolz/foolfuuka/plugins
mkdir foolz
cd foolz
git clone https://github.com/FoolCode/foolfuuka-plugin-board-statistics.git
git clone https://github.com/FoolCode/foolfuuka-plugin-dice-roll.git
git clone https://github.com/FoolCode/foolfuuka-plugin-quests.git

Set Permissions and Group for FoolFuuka Directory

sudo chmod -Rf 775 ~/foolfuuka
sudo chown -Rf www-data:www-data ~/foolfuuka

Configure Nginx to host foolfuuka

sudo nano /etc/nginx/sites-enabled/default
        root /home/ubuntu/foolfuuka/public;
        index index.php index.html index.htm;

        client_max_body_size 32M; # set to the same as max PHP upload size in HHVM
        location / {
                try_files $uri $uri/ /index.php$is_args$args;
        }

        location ~ \.php$ {
                 fastcgi_pass unix:/var/run/php5-fpm.sock;
                 fastcgi_index index.php;
                 fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
                 include fastcgi_params;
        }

Restart Nginx

sudo service nginx restart

Create the database and install Foolfuuka user accounts.

sudo mysql_secure_installation
mysql -uroot -p
  CREATE USER 'asagi'@'localhost' IDENTIFIED BY 'pass';  
  CREATE USER 'foolfuuka'@'localhost' IDENTIFIED BY 'pass';
  CREATE DATABASE asagi DEFAULT CHARACTER SET utf8mb4 DEFAULT COLLATE utf8mb4_general_ci;
  CREATE DATABASE foolfuuka DEFAULT CHARACTER SET utf8mb4 DEFAULT COLLATE utf8mb4_general_ci;
  GRANT ALL PRIVILEGES ON `asagi`. * TO 'asagi'@'localhost';
  GRANT ALL PRIVILEGES ON `asagi`. * TO 'foolfuuka'@'localhost';
  GRANT ALL PRIVILEGES ON `foolfuuka`. * TO 'foolfuuka'@'localhost';
  FLUSH PRIVILEGES;
  exit

edit MariaDB to use utf8mb4 for correct unicode processing

sudo nano /etc/mysql/my.cnf
[client]
default-character-set = utf8mb4

[mysql]
default-character-set = utf8mb4

[mysqld]
character-set-client-handshake = FALSE
character-set-server = utf8mb4
collation-server = utf8mb4_general_ci

Restart MariaDB

sudo service mysql restart

Disable transparent_hugepage

su -c 'echo never > /sys/kernel/mm/transparent_hugepage/enabled'
sudo nano /etc/default/grub

append transparent_hugepage=never to GRUB_CMDLINE_LINUX_DEFAULT, then save the file

sudo update-grub

Enable toku (Find and Uncomment the load module)

sudo nano /etc/mysql/conf.d/tokudb.cnf 

Raise file limit

sudo nano /etc/security/limits.conf
* soft nofile 63545
* hard nofile 63545
* soft nproc 63545
* hard nproc 63545

Increase mysql file limit:

sudo nano /etc/mysql/my.cnf
# Put this after [mysqld]
open-files-limit=40000
sudo service mysql stop
sudo reboot
mysql -u root -p
set global default_storage_engine=TokuDB;
SET GLOBAL log_bin_trust_function_creators = 1;

Now, Install FoolFuuka at the website using foolfuuka credentials

Install and compile Asagi from source.

cd ~
git clone https://github.com/FoolCode/asagi.git
cd asagi
mvn package assembly:single
mv target/asagi-0.4.0-SNAPSHOT-full.jar ./asagi.jar

Configure Asagi

cp asagi.json.example asagi.json
nano asagi.json

Insert the following JSON into that file:

{"settings": {
  "dumperEngine": "DumperJSON",
  "sourceEngine": "YotsubaJSON",

  "boardSettings": {
    "default": {
      "engine": "Mysql",
      "database": "asagi",
      "host": "localhost",
      "username": "asagi",
      "password": "afj8a428dauwj",
      "charset": "utf8mb4",
      "path": "/home/ubuntu/foolfuuka/public/foolfuuka/boards/",
      "useOldDirectoryStructure": false,
      "webserverGroup": "www-data",
      "thumbThreads": 3,
      "mediaThreads": 0,
      "newThreadsThreads": 3,
      "deletedThreadsThresholdPage": 8,
      "refreshDelay": 30,
      "throttleAPI": true,
      "throttleURL": "api.4cdn.org",
      "throttleMillisec": 11100,

      "threadRefreshRate": 10,
      "pageSettings": [
        {"delay": 30, "pages": [0, 1]},
        {"delay": 500, "pages": [2, 3, 4, 5, 6, 7, 8]},
        {"delay": 30, "pages": [9, 10]}
      ]
    },

    "3": {},
    "a": {},
    "aco": {},
    "adv": {},
    "an": {},
    "asp": {},
    "b": {},
    "biz": {},
    "c": {},
    "cgl": {},
    "ck": {},
    "cm": {},
    "co": {},
    "d": {},
    "diy": {},
    "e": {},
    "f": {},
    "fa": {},
    "fit": {},
    "g": {},
    "gd": {},
    "gif": {},
    "h": {},
    "hc": {},
    "hm": {},
    "his": {},
    "hr": {},
    "i": {},
    "ic": {},
    "int": {},
    "jp": {},
    "k": {},
    "lgbt": {},
    "lit": {},
    "m": {},
    "mlp": {},
    "mu": {},
    "n": {},
    "news": {},
    "o": {},
    "out": {},
    "p": {},
    "po": {},
    "pol": {},
    "qa": {},
    "qst": {},
    "r": {},
    "r9k": {},
    "s": {},
    "s4s": {},
    "sci": {},
    "soc": {},
    "sp": {},
    "t": {},
    "tg": {},
    "toy": {},
    "trash": {},
    "trv": {},
    "tv": {},
    "u": {},
    "v": {},
    "vg": {},
    "vp": {},
    "vr": {},
    "w": {},
    "wg": {},
    "wsg": {},
    "wsr": {},
    "x": {},
    "y": {}
  }
}}

Start asagi (There is a better way to do this).

Maybe we should use systemd for it...

UseParallelGC adds Parallelization to the garbage collector, which really helps for Java.

screen java -XX:+UseParallelGC -XX:+UseParallelOldGC -verbose:gc -jar asagi.jar

ATC of Desuarchive prefers these parameters, which have two customizations.

  1. For the first parameter, Java by default will use 1/4th your system memory, so you can set your own hard memory roof to reduce usage or increase it. He used 6144m.
    • If you aren't archiving more than 8 boards, this probably won't be an issue that you need to deal with.
  2. The permsize is about the permanent generation objects: depends on your heap size.
    • This setting is probably related to how many boards are on Desuarchive, but Sunako doesn't had that sort of issue.
screen java -Xmx6144m -XX:+UseParallelGC -XX:+UseParallelOldGC -verbose:gc -XX:MaxPermSize=512m -jar asagi.jar

Now, Configure asagi as the database and remove the table prefix from FoolFuuka.

Install FoolFuuka

The installer is quite self explanatory, just make sure to remove the Boards prefix.

Also make sure to install imagemagick and set the path in the config to it.

Add boards, Configure Sphinx

To configure Sphinx Search, take the following links:

http://IP/admin -> Search -> Generate Config. http://IP/admin -> Boards -> Edit -> Check Use SphinxSearch as search engine. On the Search panel, change the working directory (on Ubuntu) to the following:

/var/lib/sphinxsearch

Supplement Sphinxsearch

cd /var/run
sudo mkdir sphinxsearch
sudo chmod -Rf 777 sphinxsearch

Start Sphinxsearch

sudo nano /etc/default/sphinxsearch

Change the following line from

START=no

to

START=yes

Save and run the following command

sudo /etc/init.d/sphinxsearch start


Right-click and copy this config out of Foolfuuka and add your MySQL database information for asagi. Change the following line from

pid_file = /var/lib/sphinxsearch/searchd.pid

to

pid_file = /var/run/sphinxsearch/searchd.pid 

Then, paste this to /etc/sphinxsearch/sphinx.conf

If anything errors, kill any running instance of asagi and restart it

Build the indexes

sudo indexer *_ancient --rotate
sudo indexer *_main --rotate

Configure Crontab

sudo nano /etc/crontab
@hourly sudo -u sphinxsearch /usr/bin/indexer --rotate *_delta > /dev/null
@daily sudo -u sphinxsearch /usr/bin/indexer --rotate *_main > /dev/null
@monthly sudo -u sphinxsearch /usr/bin/indexer --rotate *_ancient > /dev/null


Adding Archives

Inject your sql file into the asagi database (assuming that your sql file is called "dump.sql"):

mysql -u root -p asagi < dump.sql

Then, in the admin panel, simply add the board that corresponds to the board you injected.

Extra configuration

Depending on which distro you use or how many posts you have, some configs might need to be changed.

Fixing thumbnail creation

Foolfuuka might not be able to generate thumbnails for images to be posted. If so, check:

$ whereis convert
convert: /usr/bin/convert /usr/share/man/man1/convert.1.gz

and make sure it is the same value in Admin panel -> general -> Imagemagick Convert

For Huge Boards: Problem with random dropped posts/threads due to stats triggers

When Desuarchive was configured with a new server using the 4plebs fork of FoolFuuka and Asagi, it was a mystery why /a/ was behind in archiving. Normal posts were being bled somehow.

But everything was running fine configuration-wise. No asagi errors. Nothing. And all other boards were great.

However, just get rid of stats triggers/procedures and it should be fine.

Asagi has triggers that update _daily and _users tables. FFuuka stats plugin reads those tables and make stats. Probably the _users table was messed up/too large for /a/, so it slows down inserts because the update runs before it.

4plebs is working on stats plugins that does it without them, since stats calc on insert isn't a smart idea.

Procedure

I just removed the calls to stats procedures. Should work but check it out first. But should leave the procedures there (commented out?) so that you can go back.

  1. Stop Asagi.
  2. Run the code below:
  3. Start Asagi.
DROP TRIGGER IF EXISTS "after_ins_a";

delimiter //

CREATE TRIGGER "after_ins_a" AFTER INSERT ON "a"
FOR EACH ROW
BEGIN
  IF NEW.op = 1 THEN
    CALL create_thread_a(NEW.num, NEW.timestamp);
  END IF;
  CALL update_thread_a(NEW.thread_num);
END//

DROP TRIGGER IF EXISTS "after_del_a"//

CREATE TRIGGER "after_del_a" AFTER DELETE ON "a"
FOR EACH ROW
BEGIN
  CALL update_thread_a(OLD.thread_num);
  IF OLD.op = 1 THEN
    CALL delete_thread_a(OLD.num);
  END IF;
  IF OLD.media_hash IS NOT NULL THEN
    CALL delete_image_a(OLD.media_id);
  END IF;
END//


delimiter ;