Flockit

Advisory file locks for Elixir using flock(2).

{:ok, lock} = Flockit.lock("/var/run/my_app.lock")
# ... only one holder at a time, across OS processes ...
:ok = Flockit.unlock(lock)

# Or scoped:
{:ok, result} = Flockit.with_lock("/var/run/my_app.lock", fn -> do_work() end)

# Shared (read) locks, timeouts and non-blocking attempts:
{:ok, lock} = Flockit.lock(path, mode: :shared, timeout: 5_000)
{:error, :eagain} = Flockit.try_lock(path)

• Never blocks the VM. Nothing ever sleeps in flock(2): waiters make non-blocking attempts and are woken by the kernel's release notifications (kqueue on macOS, inotify on Linux), with a timed retry as a backstop.
• No leaked locks. A lock is released by unlock/1, when the process that took it exits, or when it is garbage collected.
• POSIX errors. Failures come back as {:error, :enoent}, {:error, :eacces} and so on.

See the documentation for the full semantics.

Tested on Linux and macOS. Other Unix systems should work, with timed retries in place of release notifications, but are untested.

Installation

def deps do
  [
    {:flockit, "~> 1.0"}
  ]
end

A C compiler and make are needed to build the NIF.