Skip to main content

Asyncio client for Zookeeper.

Project description

Asyncio zookeeper client (aiozk)

PyPi version Build Status

Table of Contents


Have no major bugs in client/session/connection, but recipes need more test code to become more robust.

Any help and interest are welcome 😀


$ pip install aiozk

Quick Example

import asyncio
from aiozk import ZKClient

async def main():
    zk = ZKClient('localhost')
    await zk.start()
    await zk.create('/foo', data=b'bazz', ephemeral=True)
    assert b'bazz' == await zk.get_data('/foo')
    await zk.close()


You may use recipes, similar to zoonado, kazoo, and other libs:

# assuming zk is aiozk.ZKClient

# Lock
async with await'/path/to/lock').acquire():
    # ... Do some stuff ...

# Barrier
barrier ='/path/to/barrier)
await barrier.create()
await barrier.lift()
await barrier.wait()

# DoubleBarrier
double_barrier ='/path/to/double/barrier', min_participants=4)
await double_barrier.enter(timeout=0.5)
# ...  Do some stuff ...
await double_barrier.leave(timeout=0.5)

You can find full list of recipes provided by aiozk here: aiozk recipes

To understand ideas behind recipes please read this and even more recipes here. Make sure you're familiar with all recipes before doing something new by yourself, especially when it involves more than few zookeeper calls.


Don't mix different type of recipes at the same znode path. For example, creating a Lock and a DoubleBarrier object at the same path. It may cause undefined behavior 😓


NB: please ensure that you're using recent docker-compose version. You can get it by running

pip install --user -U docker-compose

Run tests

# you should have access to docker

docker-compose build

Testing approach

Most of tests are integration tests and running on real zookeeper instances. We've chosen zookeeper 3.5 version since it has an ability to dynamic reconfiguration and we're going to do all connecting/reconnecting/watches tests on zk docker cluster as this gives us the ability to restart any server and see what happens.

# first terminal: launch zookeeper cluster
docker-compose rm -fv && docker-compose build zk && docker-compose up --scale zk=7 zk_seed zk

# it will launch cluster in this terminal and remain. last lines should be like this:

zk_6       | Servers: 'server.1=;\nserver.2=;\nserver.3=;\nserver.4=;\nserver.5=;\nserver.6=;'
zk_6       | CONFIG: server.1=;
zk_6       | server.2=;
zk_6       | server.3=;
zk_6       | server.4=;
zk_6       | server.5=;
zk_6       | server.6=;
zk_6       | server.7=;
zk_6       |
zk_6       |
zk_6       | Reconfiguring...
zk_6       | ethernal loop
zk_7       | Servers: 'server.1=;\nserver.2=;\nserver.3=;\nserver.4=;\nserver.5=;\nserver.6=;\nserver.7=;'
zk_7       | CONFIG: server.1=;
zk_7       | server.2=;
zk_7       | server.3=;
zk_7       | server.4=;
zk_7       | server.5=;
zk_7       | server.6=;
zk_7       | server.7=;
zk_7       | server.8=;
zk_7       |
zk_7       |
zk_7       | Reconfiguring...
zk_7       | ethernal loop

Run tests in docker:

docker-compose run --no-deps aiozk
# last lines will be about testing results

............lot of lines ommited........
Ran 3 tests in 1.059s


Run tests locally:

# ZK_IP can be something from logs above, like: ZK_HOST=
ZK_HOST=<ZK_IP> ./venv/bin/pytest

Recipes testing

It seems that usually recipes require several things to be tested:

  • That recipe flow is working as expected
  • Timeouts: reproduce every timeout with meaningful values (timeout 0.5s and block for 0.6s)

Run some tests directly

Another way to run tests only which you are interested in quickly. Or this is useful when you run tests under the other version of python.

# Run zookeeper container
docker run -p 2181:2181 zookeeper

# Run pytest directly at the development source tree
export ZK_HOST=localhost
pytest -s --log-cli-level=DEBUG aiozk/test/


Download files

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

Files for aiozk, version 0.28.0
Filename, size File type Python version Upload date Hashes
Filename, size aiozk-0.28.0-py3-none-any.whl (74.8 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size aiozk-0.28.0.tar.gz (45.6 kB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring DigiCert DigiCert EV certificate Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page