devilbox/docs/intermediate/best-practice.rst

4.7 KiB

Best practice

If you have already operate the Devilbox, this guide is a must have. It will cover common best-practice topics as well as some tips and tricks you will want to apply.

Table of Contents

local

Move data out of Devilbox directory

One thing you should take into serious consideration is to move data such as your projects as well as persistent data of databases out of the Devilbox git directory.

The Devilbox git directory should be something that can be safely deleted and re-created without having to worry about loosing any project data. There could also be the case that you have a dedicated hard-disk to store your projects or you have your own idea about a directory structure where you want to store your projects.

Affected env variables to consider changing:

  • env_httpd_datadir
  • env_host_path_backupdir

Projects

* howto_move_projects_to_a_different_directory Follow this guide to keep your projects separated from the Devilbox git directory.

Backups

* howto_move_backups_to_a_different_directory Follow this guide to keep your backups separated from the Devilbox git directory.

Version control .env file

The .env file is ignored by git, because this is your file to customize and it should be your responsibility to make sure to backup or version controlled.

One concept you can apply here is to have a separate dotfiles git repository. This is a repository that holds all of your configuration files such as vim, bash, zsh, xinit and many more. Those files are usually stored inside this repository and then symlinked to the correct location. By having all configuration files in one place, you can see and track changes easily as well as bein able to jump back to previous configurations.

In case of the Devilbox .env file, just store this file in your repository and symlink it to the Devilbox git directiry. This way you make sure that you keep your file, even when the Devilbox git directory is deleted and you also have a means of keeping track about changes you made.

You could also go further and have several .env files available somewhere. Each of those files holds different configurations e.g. for different projects or customers.

  • env-customer1
  • env-php55
  • env-project3

You would then simply symlink one of those files to the Devilbox git directory.

Version control service config files

This will require some changes on the Devilbox and will be implemented shortly.

  • Symlink and have your own git directory
  • Separate data partition, backups

PHP project hostname settings

When configuring your PHP projects to use MySQL, PostgreSQL, Redis, Mongo and other services, make sure to set the hostname/address of each of those services to how they are defined within the Devilbox network:

As an example, if you want to access the MySQL database from within the PHP container, you do the following:

# Navigate to Devilbox git directory
host> cd path/to/devilbox

# Enter the PHP container
host> ./shell.sh

# Enter the MySQL console
php> mysql -u root -h mysql -p
mysql>

To access the MySQL database from your host operating system you would need the address to what it exposes to on your host (usually 127.0.0.1):

# Enter the MySQL console
host> mysql -u root -h 127.0.0.1 -p
mysql>

Any of your projects php files that configure MySQL as an example should point the hostname or IP address of the MySQL server to mysql:

<?php
// MySQL server connection in your project configuration
mysql_host = 'mysql';
mysql_port = '3306';
mysql_user = 'someusername';
mysql_pass = 'somepassword';
?>

work_inside_the_php_container

Timezone

The env_timezone value will affect PHP and web serverequally. It does however not affect any other official Docker container that are used within the Devilbox. This is an issue that is currently still being worked on.

Feel free to change this to any timezone you require for PHP, but keep in mind that timezone values for databases can be painful, once you want to switch to a different timezone.

A good practice is to always use UTC on databases and have your front-end application calculate the correct time for the user. This way you will be more independent of any changes.