lock_method

It’s like alias_method, but it’s lock_method!

Lets you lock a method so that only one process can call it at a time. Defaults to using lockfiles on the local file system, but can be configured to store the locks in Memcached or Redis, allowing the method locks to hold over multiple hosts.

Real-world usage

In production use at impact.brighterplanet.com and data.brighterplanet.com.

Example

require 'lock_method'
class Blog
  attr_accessor :url

  def get_latest_entries
    sleep 5
  end
  lock_method :get_latest_entries

  # used by lock_method to differentiate between instances
  def as_lock
    url
  end
end

Then you can do

my_blog.get_latest_entries => it will start...
my_blog.get_latest_entries => this will raise LockMethod::Locked if you try to run it before the other call finishes

Just in case, you can clear them

my_blog.lock_method_clear :get_latest_entries

Pays attention to method arguments

If you lock Foo.bar(*args), calling Foo.bar(:baz) will not lock out Foo.bar(:zoo).

Defining #as_lock

If you want to lock instance methods, you should define #as_lock on those instances.

Locking across hosts

If you want to lock across hosts, just use shared storage, like a remote Redis or memcached instance.

If you want to lock locally, but you’re using shared storage, just get the hostname of the locking instance into the as_lock.

Configuration (and supported cache clients)

The default is to use filesystem lockfiles, usually in /tmp/lock_method/*.

If you want to share locks among various machines, you can use a Memcached or Redis client:

LockMethod.config.storage = Memcached.new '127.0.0.1:11211'

LockMethod.config.storage = Redis.new

or this might even work…

LockMethod.config.storage = Rails.cache

See Config for the full list of supported caches.