redis_pool

A high performant Erlang pool for eredis based on erlpool

Why another redis pool

Beside the fact that the most popular project in this area eredis_pool has a low activity in the last time, it has also the following problems:

• it's using another gen_server proxy from where get's the pid of the redis connection inside the poolboy checkout. This is most probably because in the past eredis was crashing the gen_server when connection was lost and if too many crashes takes place in a short period of time the supervisor could stop. This is no longer the case as time newer versions of eredis can reconnect without crashing the gen_server.
• poolboy seems overkill for managing a pool of redis connections. Instead erlpool overhead is resumed only to one ETS counter_update operation and one lookup

Quick start

Getting all deps and compile:

rebar get-deps
rebar compile

You can define your pools inside app.config :

[
    {redis_pool, [
        {pools, [
            {mypool,[
                {size, 10},
                {host, "127.0.0.1"},
                {port, 6379},
                {database, 0},
                {reconnect_sleep, 100}
            ]}
        ]}
    ]}
].

Or you can dynamically add/remove pools at runtime using start_pool/2 and stop_pool/1 :

redis_pool:start(),
Args = [
    {size, 20},
    {host, "127.0.0.1"},
    {port, 6379}
],

redis_pool:start_pool(mypool, Args),
{ok, <<"OK">>} = redis_pool:q(mypool, [<<"SET">>, <<"foo">>, <<"bar">>]).

The supported options for a pool are:

• size : the pool size (how many connections are created)
• host : redis server address (string, default: "127.0.0.1")
• port : redis server port (integer, default: 6379)
• database : database number (integer, default to 0)
• password : database password (string, default empty password)
• reconnect_sleep: in case connection to server is lost after how many milliseconds should try to reconnect (integer, default to 100)
• connect_timeout: how many milliseconds will stay in connect until timeout (integer, default 5000)

The API is identical with the one from eredis exceptions making the transaction support:

Transactions

Transactions are implemented using pipeline, this means that you need to send the list of the commands that you want to execute in the same transaction:

redis_pool:transaction(mypool, [["SET", var1, "1"], ["LPUSH", var2, "3"], ["LPUSH", var2, "2"]]).
{ok,[<<"OK">>,<<"1">>,<<"2">>]}

Performance testing

The code is in benchmark folder. In the following test I send 300000 requests from 300 concurrent processes on a pool of 10 connections.

make bench
### 1355 ms 250000 req/sec  

You can run it yourself using make bench after you copy the load_test.erl from benchmark folder in src and compile.