dflockd-client 2.0.0

dflockd python client

dflockd-client

A Python client for dflockd — a distributed FIFO lock and counting-semaphore server.

Documentation · Changelog

Features

• Sync and async clients — pick the one that matches your runtime
• Distributed locks and counting semaphores with FIFO ordering
• Single-phase acquire and two-phase enqueue + wait
• Background lease renewal — call acquire, hold for as long as you need
• Multi-server sharding (deterministic CRC-32; matches the Go and TypeScript clients)
• TLS and shared-secret authentication
• Zero runtime dependencies; Python 3.12+

Install

pip install dflockd-client
# or
uv add dflockd-client

Quick start

A running dflockd server on 127.0.0.1:6388 is the only prerequisite.

Lock

from dflockd_client import SyncDistributedLock

with SyncDistributedLock("my-key") as lock:
    # critical section — lease auto-renews in a daemon thread
    print(f"acquired: {lock.token}")

import asyncio
from dflockd_client import AsyncDistributedLock

async def main():
    async with AsyncDistributedLock("my-key") as lock:
        print(f"acquired: {lock.token}")

asyncio.run(main())

Semaphore

Up to limit concurrent holders on the same key:

from dflockd_client import SyncDistributedSemaphore

with SyncDistributedSemaphore("pool", limit=3) as sem:
    print(f"slot: {sem.token}")

Authentication and TLS

import ssl
from dflockd_client import SyncDistributedLock

with SyncDistributedLock(
    "my-key",
    auth_token="shared-secret",
    ssl_context=ssl.create_default_context(),
) as lock:
    ...

Multi-server sharding

from dflockd_client import SyncDistributedLock

with SyncDistributedLock(
    "my-key",
    servers=[("a", 6388), ("b", 6388), ("c", 6388)],
) as lock:
    # the same key always routes to the same server
    ...

See the docs for two-phase acquisition, exception handling, low-level transport, and more.