Simple-Redis-Lock

A simple lock with Redis implemented with the pattern: https://redis.io/commands/set/. It does not aim to have a fault tolerant solution like redlock.

Motivation

This library was built to throttle the redundant queries from original server for a cache mechanism, the fault tolerance is not a matter because the worst case of losing the lock is just bringing some extra queries.

It provides deadlock prevention with an auto-release mechanism, and handles releasing conflicts with a lua script. There are only two interfaces(or public functions) for this module.

• acquire
• release

The main flow can be depicted as follows,

Eval vs EvalSha

The relesae of lock relies on a lua script to conduct an atomic check of the owner of lock. The eval command sends the script body again and again to Redis, to save some bandwidth, this library uses evalSha to run the lua script. Consequently there are some extra handlings for the load of script; basically it blindly calls the evalSha assuming the script is there and catches the NOSCRIPT error and refreshes the sha when needed.

Install

$ npm install @xavierchow/simple-redis-lock

Usage

import Redis from 'ioredis';
import { acquire } from '@xavierchow/simple-redis-lock';

const redis = new Redis();
const ttlInSec = 10;
const lock = await acquire(redis, resource, ttlInSec);
if (lock) {
  try {
    // some processing here
  } finally {
    await lock.release();
  }
}

Development

# start redis with docker
$ ./dockers.sh

# run the test
$ npm test

License

MIT