Skip to main content
CI Test Status

Overview

Portalocker is a library to provide an easy API to file locking.

Portalocker requires Python 3.10 or later.

An important detail to note is that on Linux and Unix systems the locks are advisory by default. By specifying the -o mand option to the mount command it is possible to enable mandatory file locking on Linux. This is generally not recommended however. For more information about the subject:

Windows and the pywin32 dependency

Since version 4.0.0, pywin32 is no longer installed automatically on Windows. Exclusive locks work out of the box without any extra dependencies using the built-in msvcrt module. Shared locks (LockFlags.SHARED) on Windows require the optional pywin32 dependency, which can be installed through the win32 extra:

pip install "portalocker[win32]"

Attempting to acquire a shared lock on Windows without pywin32 raises an ImportError explaining this requirement.

The module is currently maintained by Rick van Hattem <Wolph@wol.ph>. The project resides at https://github.com/WoLpH/portalocker . Bugs and feature requests can be submitted there. Patches are also very welcome.

Security contact information

To report a security vulnerability, please use the Tidelift security contact. Tidelift will coordinate the fix and disclosure.

PidFileLock contexts

The default context is inspection-oriented: it enters even when another process owns the lock and returns that process’s PID. A None value means this process acquired the lock:

import portalocker

with portalocker.PidFileLock('worker.pid') as holder_pid:
    if holder_pid is None:
        run_singleton_worker()
    else:
        print(f'worker already running as PID {holder_pid}')

Use fail_closed() when the protected body must only run after acquisition:

import portalocker

try:
    with portalocker.PidFileLock('worker.pid').fail_closed():
        run_singleton_worker()
except portalocker.AlreadyLocked as exc:
    print(f'worker already running as PID {exc.holder_pid}')

Redis Locks

This library now features a lock based on Redis which allows for locks across multiple threads, processes and even distributed locks across multiple computers.

It is an extremely reliable Redis lock that is based on pubsub.

As opposed to most Redis locking systems based on key/value pairs, this locking method is based on the pubsub system. The big advantage is that if the connection gets killed due to network issues, crashing processes or otherwise, it will still immediately unlock instead of waiting for a lock timeout.

First make sure you have everything installed correctly:

pip install "portalocker[redis]"

Usage is really easy:

import portalocker

lock = portalocker.RedisLock('some_lock_channel_name')

with lock:
    print('do something here')

Shared locks allow multiple readers while remaining mutually exclusive with writers:

read_lock = portalocker.RedisLock(
    'some_lock_channel_name',
    flags=portalocker.LockFlags.SHARED,
)

with read_lock:
    print('read shared state')

Waiting writers prevent new readers from entering, so a continuous stream of readers cannot starve an exclusive lock. New shared locks also interoperate with older portalocker clients: legacy Redis locks are treated as exclusive.

The API is essentially identical to the other Lock classes so in addition to the with statement you can also use lock.acquire(...).

The normal test suite uses fakeredis. To optionally repeat the Redis tests against a locally running server:

redis-server
tox -e redis-live

Set REDIS_HOST or REDIS_PORT when the server does not use localhost:6379. The redis-live environment fails instead of skipping when it cannot connect.

Upgrading from 3.x to 4.0.0

Version 4.0.0 raised the minimum Python to 3.10, stopped installing pywin32 on Windows by default, and changed what str() returns for POSIX lock exceptions. The migration guide walks through every change that can require an edit, with the reason and the fix for each.

Tips

On some networked filesystems it might be needed to force a os.fsync() before closing the file so it’s actually written before another client reads the file. Effectively this comes down to:

with portalocker.Lock('some_file', 'rb+', timeout=60) as fh:
    # do what you need to do
    ...

    # flush and sync to filesystem
    fh.flush()
    os.fsync(fh.fileno())

Examples

To make sure your cache generation scripts don’t race, use the Lock class:

>>> import portalocker
>>> with portalocker.Lock('somefile', timeout=1) as fh:
...     print('writing some stuff to my cache...', file=fh)

To customize the opening and locking a manual approach is also possible:

>>> import portalocker
>>> file = open('somefile', 'r+')
>>> portalocker.lock(file, portalocker.LockFlags.EXCLUSIVE)
>>> _ = file.seek(12)
>>> _ = file.write('foo')

Explicitly unlocking is not needed in most cases but omitting it has been known to cause issues: https://github.com/AzureAD/microsoft-authentication-extensions-for-python/issues/42#issuecomment-601108266

If needed, it can be done through:

>>> portalocker.unlock(file)
>>> file.close()

Do note that your data might still be in a buffer so it is possible that your data is not available until you flush() or close().

To create a cross platform bounded semaphore across multiple processes you can use the NamedBoundedSemaphore class which functions somewhat similar to threading.BoundedSemaphore. The name is what ties the participating processes to the same set of slots, so give it one you control:

>>> import portalocker
>>> n = 2
>>> timeout = 0.1
>>> name = 'readme_example'
>>> semaphore_a = portalocker.NamedBoundedSemaphore(
...     n, name=name, timeout=timeout)
>>> semaphore_b = portalocker.NamedBoundedSemaphore(
...     n, name=name, timeout=timeout)
>>> semaphore_c = portalocker.NamedBoundedSemaphore(
...     n, name=name, timeout=timeout)
>>> semaphore_a.acquire()
<portalocker.utils.Lock object at ...>
>>> semaphore_b.acquire()
<portalocker.utils.Lock object at ...>
>>> semaphore_c.acquire()
Traceback (most recent call last):
  ...
portalocker.exceptions.AlreadyLocked
>>> semaphore_a.release()
>>> semaphore_b.release()

More examples can be found in the tests.

Versioning

This library follows Semantic Versioning.

Changelog

Every release has a git tag with a commit message for the tag explaining what was added and/or changed. The list of tags/releases including the commit messages can be found here: https://github.com/WoLpH/portalocker/releases

License

See the LICENSE file.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

portalocker-4.1.0.tar.gz (97.8 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

portalocker-4.1.0-py3-none-any.whl (65.9 kB view details)

Uploaded Python 3

File details

Details for the file portalocker-4.1.0.tar.gz.

File metadata

  • Download URL: portalocker-4.1.0.tar.gz
  • Upload date:
  • Size: 97.8 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for portalocker-4.1.0.tar.gz
Algorithm Hash digest
SHA256 91d0ff02d6f5f9a2dbf1ef1367ddb05b83136b2676707a1b2c8746f1f0dcd992
MD5 906056cca41cddfff76d49b12a9797c8
BLAKE2b-256 45f724b2821a077e0d85197a8c4093a4f71f78713d0a6f216c3f7cc90f7a5cc4

See more details on using hashes here.

Provenance

The following attestation bundles were made for portalocker-4.1.0.tar.gz:

Publisher: publish.yml on wolph/portalocker

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file portalocker-4.1.0-py3-none-any.whl.

File metadata

  • Download URL: portalocker-4.1.0-py3-none-any.whl
  • Upload date:
  • Size: 65.9 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for portalocker-4.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 d985a430d265adf31adf12bc0bf3501aea59efc495e9104c057e5dfb7394c226
MD5 6408631ee9ea70b27a8b0a874e246d55
BLAKE2b-256 6191288c883303be067c1648f1e63ea38e5f8eb5ab7123fd3a9a7366148e58b7

See more details on using hashes here.

Provenance

The following attestation bundles were made for portalocker-4.1.0-py3-none-any.whl:

Publisher: publish.yml on wolph/portalocker

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page