lock_and_cache
Lock and cache using redis!
Most caching libraries don't do locking, meaning that >1 process can be calculating a cached value at the same time. Since you presumably cache things because they cost CPU, database reads, or money, doesn't it make sense to lock while caching?
Quickstart
const lnc = let cache = lock_and_cache // standalone mode { return cache} // wrap modeconst getStockQuote = cache // custom options w immediate closing// options are passed thru to cache manager which passes thru to backendsconst customCache = caches: cache
Install
npm install --save lock_and_cache
Theory
lock_and_cache...
- returns cached value (if exists)
- acquires a lock
- returns cached value (just in case it was calculated while we were waiting for a lock)
- calculates and caches the value
- releases the lock
- returns the value
As you can see, most caching libraries only take care of (1) and (4) (well, and (5) of course).
Setup
Just setting REDIS_URL in your environment is enough.
REDIS_URL=redis://redis:6379/2
If you want to put the cache and the locks in different places (which is useful if you want to invalidate the cache without messing with any locks currently held), you can do:
CACHE_URL=redis://redis:6379/2 LOCK_URL=redis://redis:6379/3
Distributed locking
Distributed locking is supported and uses redlock.
const lockAndCache = const cache = lockAndCache
All clients you specify will be used for locking, but only the first will be used for caching.
API
lockAndCache (mixed key, number ttl, function work)
- returns Promise
- key can be a number, string, boolean, object, etc. It will be passed to JSON.stringify.
- ttl is the time-to-live for the cached value in seconds
- work is the function that will be executed if the value is not cached. It can return a promise.
lockAndCache.configure (mixed client(s)[, object options])
- returns a lockAndCache function with the given configuration
- client(s) can be a connection url, a redis options object, or an array of connection urls and/or options objects
- options get passed to redlock
Global configuration
The LOCK_DRIFT_FACTOR
, LOCK_RETRY_COUNT
, and LOCK_RETRY_DELAY
environment
variables can be set to configure the global lockAndCache function's locking
behavior. You probably don't need this unless the jobs you are caching are
either very fast or very slow. By default, the global lockAndCache function
will retry every 100ms for one hour using the standard drift factor.
Contributing
Please send me your pull requests!