https://github.com/zendesk/active_record_host_pool

Connect to multiple databases using one ActiveRecord connection
https://github.com/zendesk/active_record_host_pool

Last synced: about 1 year ago
JSON representation

Connect to multiple databases using one ActiveRecord connection

• Host: GitHub
• URL: https://github.com/zendesk/active_record_host_pool
• Owner: zendesk
• License: mit
• Created: 2011-04-06T16:18:31.000Z (about 15 years ago)
• Default Branch: main
• Last Pushed: 2025-02-03T20:46:29.000Z (over 1 year ago)
• Last Synced: 2025-03-31T07:06:58.552Z (over 1 year ago)
• Language: Ruby
• Homepage:
• Size: 550 KB
• Stars: 16
• Watchers: 414
• Forks: 6
• Open Issues: 3
• Metadata Files:
  • Readme: Readme.md
  • Changelog: Changelog.md
  • License: MIT-LICENSE
  • Codeowners: .github/CODEOWNERS

Awesome Lists containing this project

README

          
[![Build Status](https://github.com/zendesk/active_record_host_pool/workflows/CI/badge.svg)](https://github.com/zendesk/active_record_host_pool/actions?query=workflow%3ACI)

# ActiveRecord host pooling

This gem allows for one ActiveRecord connection to be used to connect to multiple databases on a server.

It accomplishes this by calling select_db() as necessary to switch databases between database calls.

## How Connections Are Pooled

ARHP creates separate connection pools based on the pool key.

The pool key is defined as:

`host / port / socket / username / replica`

Therefore two databases with identical host, port, socket, username, and replica status will share a connection pool.

If any part (host, port, etc.) of the pool key differ, two databases will _not_ share a connection pool.

`replica` in the pool key is a boolean indicating if the database is a replica/reader (true) or writer database (false).

Below, `test_pool_1` and `test_pool_2` have identical host, username, socket, and replica status but the port information differs.

Here the database configurations are formatted as a table to give a visual example:

|          |  test_pool_1   |  test_pool_2   |

|----------|----------------|----------------|

| host     | 127.0.0.1      | 127.0.0.1      |

| port     |                | 3306           |

| socket   |                |                |

| username | root           | root           |

| replica  | false          | false          |

The configuration items must be explicitly defined or they will be blank in the pool key.

Configurations with matching _implicit_ items but differing _explicit_ items will create separate pools.

e.g. `test_pool_1` will default to port 3306 but because it is not explicitly defined it will not share a pool with `test_pool_2`

ARHP will therefore create the following pool keys:

```

test_pool_1 => 127.0.0.1///root/false

test_pool_2 => 127.0.0.1/3306//root/false

```

## Support

For now, the only backend known to work is MySQL, with the mysql2 or activerecord-trilogy-adapter gem. When using the activerecord-trilogy-adapter ensure that the transitive dependency Trilogy is v2.5.0+.

Postgres, from an informal reading of the docs, will never support the concept of one server connection sharing multiple dbs.

## Installation

    $ gem install active_record_host_pool

and make sure to require 'active\_record\_host\_pool' in some way.

## Testing

You need a local user called 'john-doe'.

    mysql -uroot

    CREATE USER 'john-doe'@'localhost';

    GRANT SELECT,INSERT,UPDATE,DELETE,CREATE,DROP,INDEX  ON *.* TO 'john-doe'@'localhost';

    FLUSH PRIVILEGES;

With mysql running locally, run

    BUNDLE_GEMFILE=gemfiles/rails6.1.gemfile bundle exec rake test

 Or

    BUNDLE_GEMFILE=gemfiles/rails6.1.gemfile ruby test/test_arhp.rb --seed 19911 --verbose

## Copyright

Copyright (c) 2011 Zendesk. See MIT-LICENSE for details.

## Authors

Ben Osheroff ,

Mick Staugaard