question-mark
Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Creating an AioClient locks up in distributed 1.25.3

See original GitHub issue

This short reproducer works in distributed 1.25.2, but locks up in 1.25.3. This behavior is triggered in https://github.com/LiberTEM/LiberTEM/ when creating a dask cluster and leads to a hard freeze of the application.

Update: The behavior is observed both on Windows 10 and on Linux.

import asyncio

from dask import distributed as dd
from distributed.asyncio import AioClient


async def async_main():
    cluster_kwargs = {
        "threads_per_worker": 1,
        "asynchronous": True,
        "n_workers": 2
    }
    client_kwargs = {}
    print("starting cluster...")
    cluster = dd.LocalCluster(**(cluster_kwargs or {}))
    print("cluster started.")
    print("creating client...")
    client = await AioClient(cluster, **(client_kwargs or {}))
    print("success")


if __name__ == '__main__':
    loop = asyncio.get_event_loop()
    loop.run_until_complete(async_main())

Expected behavior: It continues working the same in all version past 1.25.2.

Issue Analytics

  • State:closed
  • Created 5 years ago
  • Comments:17 (9 by maintainers)

github_iconTop GitHub Comments

2reactions
mrocklincommented, Feb 20, 2019

I’m curious, can you just use a normal Client? Tornado and Asyncio have merged to the point where I’m starting to suspect that AioClient may no longer be worth the maintenance effort.

On Wed, Feb 20, 2019 at 4:06 AM Dieter Weber notifications@github.com wrote:

This short reproducer works in distributed 1.25.2, but locks up in 1.25.3. This behavior is triggered in https://github.com/LiberTEM/LiberTEM/ when creating a dask cluster and leads to a hard freeze of the application.

import asyncio from dask import distributed as ddfrom distributed.asyncio import AioClient

async def async_main(): cluster_kwargs = { “threads_per_worker”: 1, “asynchronous”: True, “n_workers”: 2 } client_kwargs = {} cluster = dd.LocalCluster(**(cluster_kwargs or {})) client = await AioClient(cluster, **(client_kwargs or {})) print(“success”)

if name == ‘main’: loop = asyncio.get_event_loop() loop.run_until_complete(async_main())

Expected behavior: It continues working the same in all version past 1.25.2.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/dask/distributed/issues/2537, or mute the thread https://github.com/notifications/unsubscribe-auth/AASszE3vijKQum5bQGTaNuOXO6oVXCLEks5vPTpMgaJpZM4bFEYb .

0reactions
kszucscommented, Feb 22, 2019

Yeah, I suggest pinning tornado 5 and deprecating AioClient as a wrapper for Client(asynchronous=True).

Read more comments on GitHub >

github_iconTop Results From Across the Web

Changelog — Dask.distributed 2022.12.1 documentation
Note. This release changes the default scheduling mode to use queuing. This will significantly reduce cluster memory use in most cases, and generally ......
Read more >
Docker Compose release notes | Docker Documentation
Fixed a bug where working with containers created with a version of Compose earlier than 1.23.0 would cause unexpected crashes.
Read more >
Google Cloud release notes | Documentation
Create and use a list of preferred load-balancing policies. With this feature, if your first preferred policy can't be used by a particular...
Read more >
Configuration and schema reference Red Hat Directory Server ...
It is possible for an attacker to send such a large packet size that it crashes Directory Server or ties it up indefinitely...
Read more >
Kubernetes In-Tree to CSI Volume Migration Status Update
The solution we came up with was CSI migration: a feature that translates ... If you are a Kubernetes distribution maintainer, make sure...
Read more >

github_iconTop Related Medium Post

No results found

github_iconTop Related StackOverflow Question

No results found

github_iconTroubleshoot Live Code

Lightrun enables developers to add logs, metrics and snapshots to live code - no restarts or redeploys required.
Start Free

github_iconTop Related Reddit Thread

No results found

github_iconTop Related Hackernoon Post

No results found

github_iconTop Related Tweet

No results found

github_iconTop Related Dev.to Post

No results found

github_iconTop Related Hashnode Post

No results found