webtorrent/webtorrent

Over 75% CPU usage and most of the run time is spent in epoll_pwait

Open

#1926 opened on Oct 5, 2020

View on GitHub
 (3 comments) (0 reactions) (0 assignees)JavaScript (30,002 stars) (2,748 forks)batch import
acceptedarea/perfhelp wanted

Description

What version of this package are you using? WebTorrent 0.108.6

What operating system, Node.js, and npm version? OS: Node:12 Docker image Node: v14.12.0 npm: 6.14.8

What happened? Running the example provided in the "Getting Started" section inside a docker container will cause the CPU usage to jump to around 50%. and if a web server is created for the torrent (torrent.createServer(port)), then the usage increases further to around 75%. this behavior was observed when running the container on multiple machines running Windows and Linux. however, if the same code is run on a windows machine (not inside a container) then no abnormal CPU usage is seen.

I've done a profile using the node profiler inside the container, and it seems that most of the time (around 74%) is spent in epoll_pwait which can be seen here.

with the current behavior, Webtorrent is unusable in Docker on relatively under power devices since it will literally lag the system.

PS: here is the code I've used for testing just in case.

let WebTorrent = require('webtorrent')

let client = new WebTorrent()
let port = 40000;
// Sintel, a free, Creative Commons movie
let torrentId = 'magnet:?xt=urn:btih:08ada5a7a6183aae1e09d831df6748d566095a10&dn=Sintel&tr=udp%3A%2F%2Fexplodie.org%3A6969&tr=udp%3A%2F%2Ftracker.coppersurfer.tk%3A6969&tr=udp%3A%2F%2Ftracker.empire-js.us%3A1337&tr=udp%3A%2F%2Ftracker.leechers-paradise.org%3A6969&tr=udp%3A%2F%2Ftracker.opentrackr.org%3A1337&tr=wss%3A%2F%2Ftracker.btorrent.xyz&tr=wss%3A%2F%2Ftracker.fastcast.nz&tr=wss%3A%2F%2Ftracker.openwebtorrent.com&ws=https%3A%2F%2Fwebtorrent.io%2Ftorrents%2F&xs=https%3A%2F%2Fwebtorrent.io%2Ftorrents%2Fsintel.torrent'

client.add(torrentId, (torrent) => {
    let server = torrent.createServer();
    server.listen(port);
    console.log("ready!")
})

Contributor guide

Tech stack
javascriptnodejs
Domain
performance
Issue type
performance
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
4
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-3 hours
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
clear
Prerequisites
Node.js basicsDocker experienceNode.js profiling
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
30
Research direction
Investigate CPU usage on Docker by examining the event loop and libuv's epoll wait usage. The provided profile shows 74% time in epoll pwait, which may indicate excessive polling or a missing sleep. Review node's fs.watch or timer implementations. Look at WebTorrent's source for any setInterval or synchronous loops. Consider if the issue is due to Docker's eventfd behavior or a missing uv unref. The pastebin link (https://pastebin.com/BEe9Lqra) contains the full profile. Focus on the ` poll` function in the core node files.