PGPool

PGPool is a PosgreSQL client that automatically uses connection pools and handles reconnections in case of errors. PGPool also optimizes all of your statements, by preparing them and caching them for you under the hood.

It uses:

• epgsql as the PostgreSQL driver.
• poolboy as pooling library.

Install

For Elixir

Add it to your deps:

defp deps do
  [{:pgpool, "~> 2.1"}]
end

Ensure that pgpool is started with your application, for example by adding it to the list of your application's extra_applications:

def application do
  [
    extra_applications: [:logger, :pgpool]
  ]
end

For Erlang

If you're using rebar3, add pgpool as a dependency in your project's rebar.config file:

{pgpool, {git, "git://github.com/ostinelli/pgpool.git", {tag, "2.1.0"}}}

Or, if you're using Hex.pm as package manager (with the rebar3_hex plugin):

{pgpool, "2.1.0"}

Ensure that pgpool is started with your application, for example by adding it in your .app file to the list of applications:

{application, my_app, [
    %% ...
    {applications, [
        kernel,
        stdlib,
        sasl,
        pgpool,
        %% ...
    ]},
    %% ...
]}.

Usage

Since pgpool is written in Erlang, the example code here below is in Erlang. Thanks to Elixir interoperability, the equivalent code in Elixir is straightforward.

Specify Databases

Databases can be set in the environment variable pgpool. You're probably best off using an application configuration file (in releases, sys.config):

{pgpool, [
  {databases, [
    {db1_name, [
      {pool, [
        %% poolboy options <https://github.com/devinus/poolboy>
        %% The `name` and `worker_module` options here will be ignored.
        {size, 10},        %% maximum pool size
        {max_overflow, 5}, %% maximum number of additional workers created if pool is empty
        {strategy, lifo}   %% can be lifo or fifo (default is lifo)
      ]},
      {connection, [
        {host, "localhost"},
        {user, "postgres"},
        {pass, ""},
        {options, [
          %% epgsql connect_options() <https://github.com/epgsql/epgsql>
          {port, 5432},
          {ssl, false},
          {database, "db1"}
        ]}
      ]}
    ]},

    {db2_name, [
      {pool, [
        {size, 10},
        {max_overflow, 20},
        {strategy, lifo}
      ]},
      {connection, [
        {host, "localhost"},
        {user, "postgres"},
        {pass, ""},
        {options, [
          {port, 5432},
          {ssl, false},
          {database, "db2"}
        ]}
      ]}
    ]}
  ]}
]}

Custom types

-type pgpool_query_option() :: no_wait.

no_wait makes the query call return immediately if there are no available connections in the pool. This allows you to improve your application's flow control by rejecting external calls if your database is unable to handle the load, thus preventing a system overflow.

Queries

Please refer to epgsql 3.4 README for how to perform queries. Currently, PGPool supports the following.

Simple Query

pgpool:squery(DatabaseName, Sql) ->
  pgpool:squery(DatabaseName, Sql, []).

pgpool:squery(DatabaseName, Sql) ->
  Result

Types:
  DatabaseName = atom()
  Sql = string() | iodata()
  Options = [pgpool_query_option()]
  Result = {ok, Count} | {ok, Count, Rows} | {error, no_connection | no_available_connections}
  Count = non_neg_integer()
  Rows = (see epgsql for more details)

For example:

pgpool:squery(db1_name, "SELECT * FROM users;").

Simple queries cannot be optimized by PGPool since they cannot be prepared. If you want to optimize and cache your queries, consider using equery/3,4 or batch/2 instead.

Extended Query

pgpool:equery(DatabaseName, Statement, Params) ->
  pgpool:equery(DatabaseName, Statement, Params, []).

pgpool:equery(DatabaseName, Statement, Params, Options) ->
  Result

Types:
  DatabaseName = atom()
  Statement = string()
  Params = list()
  Options = [pgpool_query_option()]
  Result = {ok, Count} | {ok, Count, Rows} | {error, no_connection | no_available_connections}
  Count = non_neg_integer()
  Rows = (see epgsql for more details)

For example:

pgpool:equery(db1_name, "SELECT * FROM users WHERE id = $1;", [3]).

PGPool will prepare your statements and cache them for you, which results in considerable speed improvements. If you use a lot of different statements, consider memory usage because the statements are not garbage collected.

Batch Queries

To execute a batch:

pgpool:batch(DatabaseName, StatementsWithParams) ->
  pgpool:batch(DatabaseName, StatementsWithParams, Options).

pgpool:batch(DatabaseName, StatementsWithParams) ->
  Result

Types:
  DatabaseName = atom()
  StatementsWithParams = [{Statement, Params}]
  Statement = string()
  Params = list()
  Options = [pgpool_query_option()]
  Result = [{ok, Count} | {ok, Count, Rows}] | {error, no_connection | no_available_connections}
  Count = non_neg_integer()
  Rows = (see epgsql for more details)

For example:

S = "INSERT INTO users (name) VALUES ($1);",

[{ok, 1}, {ok, 1}] = pgpool:batch(db1_name, [
  {S, ["Hedy"]},
  {S, ["Roberto"]}
]).

PGPool will prepare your statements and cache them for you, which results in considerable speed improvements. If you use a lot of different statements, consider memory usage because the statements are not garbage collected.

Contributing

So you want to contribute? That's great! Please follow the guidelines below. It will make it easier to get merged in.

Before implementing a new feature, please submit a ticket to discuss what you intend to do. Your feature might already be in the works, or an alternative implementation might have already been discussed.

Do not commit to master in your fork. Provide a clean branch without merge commits. Every pull request should have its own topic branch. In this way, every additional adjustments to the original pull request might be done easily, and squashed with git rebase -i. The updated branch will be visible in the same pull request, so there will be no need to open new pull requests when there are changes to be applied.

Ensure that proper testing is included. To run PGPool tests, you need to create the database pgpool_test for user postgres with no password, and then simply run from the project's root directory:

$ make test