Dokploy•7mo ago

Headless Chrome Timing Out on Dokploy Deployment

I have an issue getting a headless Chrome browser to run in my Ruby on Rails app deployed via Dokploy. The browser times out on all requests when started from the web server, but it works perfectly when launched from the Rails console in the Terminal (via Dokploy UI). The goal is to fetch a remote URL and extract its content. My Docker setup is solid, as the app runs smoothly when built and executed with plain Docker, on Coolify (a competitor), and on my local development machine. The problem appears to be specific to the Dokploy deployment. The only feedback I get from Dokploy logs is a series of messages like:

[1] ! reaped unknown child process pid=1709 status=pid 1709 exit 0
[1] ! reaped unknown child process pid=1805 status=pid 1805 exit 0
(lots more)

[1] ! reaped unknown child process pid=1709 status=pid 1709 exit 0
[1] ! reaped unknown child process pid=1805 status=pid 1805 exit 0
(lots more)

I’m not sure what’s causing this. I'm not even sure it's related to Dokploy. However I'm at my wits end. Anybody have any ideas? Dockerfile: https://pastebin.com/GywVfPT7 Chrome (error) messages when starting (successfully) from Rails console: https://pastebin.com/R7xfT21a

Pastebin

syntax=docker/dockerfile:1# check=error=true# This Dockerfile is ...

Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.

Pastebin

rails@b26c20becbfa:/rails$ /usr/bin/google-chrome --disable-http2 -...

Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.

26 Replies

Siumauricio•7mo ago

Hi the docker image build was successful? are you deploying in dokploy server or remote?

johannes0910OP•7mo ago

Heyo, Yes, the build works fine and my app runs well. I deploy it to my VPS where I run Dokploy. It's the same image I ran on Coolify & on bare metal. Just on Dokploy it fails to execute chrome.

Siumauricio•7mo ago

hmm what did you did on coolify? like any special setting or something?

johannes0910OP•7mo ago

exactly the same thing ... I registered my SCM, told it which Dockerfile to use and then I pressed the Deploy button. No special settings, except the usual env variables.

Siumauricio•7mo ago

can you enter to the server and do docker service ls

johannes0910OP•7mo ago

docker service ls
ID             NAME                        MODE         REPLICAS   IMAGE                              PORTS
ygi3v52pb24t   directus-cms-rails-b2ac39   replicated   0/0        directus-cms-rails-b2ac39:latest
y93em5q9kvgu   directus-database-45e7db    replicated   1/1        pgvector/pgvector:pg16
nllwbpe80bbe   dokploy                     replicated   1/1        dokploy/dokploy:latest
zxc6p2w58m5o   dokploy-postgres            replicated   1/1        postgres:16
ycri0zszrb28   dokploy-redis               replicated   1/1        redis:7
7u3feoorhvz0   dokploy-traefik             replicated   1/1        traefik:v3.1.2
moasjrpfwmi8   firerails-postgres-9d7161   replicated   1/1        pgvector/pgvector:pg16
jq9z0vqrpse0   firerails-rails-18eb52      replicated   1/1        firerails-rails-18eb52:latest
do21gulbplh5   firerails-redis-95d863      replicated   1/1        redis:7

docker service ls
ID             NAME                        MODE         REPLICAS   IMAGE                              PORTS
ygi3v52pb24t   directus-cms-rails-b2ac39   replicated   0/0        directus-cms-rails-b2ac39:latest
y93em5q9kvgu   directus-database-45e7db    replicated   1/1        pgvector/pgvector:pg16
nllwbpe80bbe   dokploy                     replicated   1/1        dokploy/dokploy:latest
zxc6p2w58m5o   dokploy-postgres            replicated   1/1        postgres:16
ycri0zszrb28   dokploy-redis               replicated   1/1        redis:7
7u3feoorhvz0   dokploy-traefik             replicated   1/1        traefik:v3.1.2
moasjrpfwmi8   firerails-postgres-9d7161   replicated   1/1        pgvector/pgvector:pg16
jq9z0vqrpse0   firerails-rails-18eb52      replicated   1/1        firerails-rails-18eb52:latest
do21gulbplh5   firerails-redis-95d863      replicated   1/1        redis:7

` This is what I am expecting. The firerails-rails image hosts chrome

Siumauricio•7mo ago

I guess the firerails-rails-18eb52 is the service which is giving the issues right? if so can you do docker service logs firerails-rails-18eb52

johannes0910OP•7mo ago

this is the code which times out: https://pastebin.com/aAps4phn ... nothing special to see here (I think)

Pastebin

Ruby Script for Web Content Extraction## loads the page in puppet...

Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.

Siumauricio•7mo ago

can you print the logs?

johannes0910OP•7mo ago

not sure ... I think the logs are old? That's the last couple of lines:

firerails-rails-18eb52.1.3cu2vgi83bm9@dokploy    | 2024-10-27 05:59:57 +0000 [DEBUG]:   JobRun Update (6.2ms)  UPDATE "job_runs" SET "end_date" = $1, "result" = $2, "updated_at" = $3 WHERE "job_runs"."id" = $4  [["end_date", "2024-10-27 05:59:56.980209"], ["result", 1], ["updated_at", "2024-10-27 05:59:56.987332"], ["id", 395161]]
firerails-rails-18eb52.1.3cu2vgi83bm9@dokploy    | 2024-10-27 05:59:57 +0000 [DEBUG]:   TRANSACTION (2.4ms)  COMMIT
firerails-rails-18eb52.1.3cu2vgi83bm9@dokploy    | [1] === puma shutdown: 2024-10-27 06:00:07 +0000 ===
firerails-rails-18eb52.1.3cu2vgi83bm9@dokploy    | [1] - Goodbye!
firerails-rails-18eb52.1.3cu2vgi83bm9@dokploy    | [1] - Gracefully shutting down workers...

firerails-rails-18eb52.1.3cu2vgi83bm9@dokploy    | 2024-10-27 05:59:57 +0000 [DEBUG]:   JobRun Update (6.2ms)  UPDATE "job_runs" SET "end_date" = $1, "result" = $2, "updated_at" = $3 WHERE "job_runs"."id" = $4  [["end_date", "2024-10-27 05:59:56.980209"], ["result", 1], ["updated_at", "2024-10-27 05:59:56.987332"], ["id", 395161]]
firerails-rails-18eb52.1.3cu2vgi83bm9@dokploy    | 2024-10-27 05:59:57 +0000 [DEBUG]:   TRANSACTION (2.4ms)  COMMIT
firerails-rails-18eb52.1.3cu2vgi83bm9@dokploy    | [1] === puma shutdown: 2024-10-27 06:00:07 +0000 ===
firerails-rails-18eb52.1.3cu2vgi83bm9@dokploy    | [1] - Goodbye!
firerails-rails-18eb52.1.3cu2vgi83bm9@dokploy    | [1] - Gracefully shutting down workers...

` It's 08:37 now

Siumauricio•7mo ago

hmm i remember someone created a discussion about puppeter, I don't know if could be helpfull https://github.com/Dokploy/dokploy/discussions/372

GitHub

VPS performance issue ? · Dokploy dokploy · Discussion #372

I noted something weird regarding the additional processing head dokploy/containers add on top of vanilla VPS deploy. Here are my 2 identical contabo VPS ( 4vCPU / 6GB RAM / 200 GB ). On first one ...

Siumauricio•7mo ago

did you assign something in advanced settings?

johannes0910OP•7mo ago

apart from some volumes, no it's all default

Siumauricio•7mo ago

how do you access it through a url or how?

johannes0910OP•7mo ago

From the issue: "Dokploy doesn't do anything than just building the image and running in to a container, so is basically the same thing you run the commands manually in your server, but dokploy automatize that process" That's what I assume too and that's why I don't understand why it doesn't work. It doesn't help me though. My server has 16gb ram and lots of cores. It's not a resource issue. Besides, it works in rails console??? 1) Working: - Dokploy terminal - bin/rails console - ExtractContent.new.fetch_content("https://www.google.com") 2) Not working: - The same ExtractContent.new.fetch_content called via a controller action or background job Both are running in the same container.

Siumauricio•7mo ago

It is correct what you mention about the issue, it is 1:1 construction and execution, it should work, you already had docker installed previously on that machine or it is a new server? And how do you trigger that action? via http or what?

johannes0910OP•7mo ago

yes. Controller Actions are HTTP Endpoints, such as /foo Alternatively I start a background job that gets processed in the same process as the webserver. Both fail in production on Dokploy Sample Code that works, in Dev:

    def test1
        url = "https://www.heise.de"
        content = ExtractContent.new.fetch_content(url)

        render :plain => content
    end

    def test1
        url = "https://www.heise.de"
        content = ExtractContent.new.fetch_content(url)

        render :plain => content
    end

` Very simple ... it fails / hangs on my VPS

Siumauricio•7mo ago

how did you set the domain? Because I assume you are triggering the container via some url.

johannes0910OP•7mo ago

its https://mydomain/test basically. The app works fine.

Siumauricio•7mo ago

hmm have you tried to search something about this: reaped unknown child process pid=1709 status=pid 1709 exit 0 this means there are zombie proceses because I see this on this issue https://github.com/puma/puma/discussions/3420 https://github.com/puma/puma/issues/3313

GitHub

Seeing multiple "reaped unknown child process" messages · puma puma...

When we run application we are seeing multiple logs like "! reaped unknown child process pid=34 status=pid 34 exit 0" is there a way we can disable this log in our application ? To Reprod...

GitHub

Puma cluster not reaping child processes with Puma 6.4.1 · Issue #3...

Describe the bug We have a separate fleet of Puma workers to handle ActionCable, and since upgrading to 6.4.1 we have seen a significant increase of unhealthy pods and can't alloc thread error ...

Siumauricio•7mo ago

I'm not very familiar with rails tbh, I remember someone asked me something similar to that and in the end it was something that had to be done internally in the application. also you already had docker installed previously on that machine or it is a new server?

johannes0910OP•7mo ago

it's a brand new Dokploy-online hetzner VPS the Puma threads don't help me, because I don't fully grasp how things work internally. Rails usually works like magic... until it doesn't. Updating to Rails 8 was another thing I wanted to try - but then again, it works on my machine (tm) all I can say now is that my simple demo code outputs 10 warnings(?) like this: [1] ! reaped unknown child process pid=218 status=pid 218 exit 0

Siumauricio•7mo ago

I would like to help you, I think this is a configuration or application problem, do you have a way to reproduce it and with the steps to follow? I am not very familiar with rails. I remember that in go I did something similar that created dead processes, and with some configuration I made that it did not continue happening.

johannes0910OP•7mo ago

I have to set up a demo system for that, this will take some time

Siumauricio•7mo ago

Ok, let me know!

johannes0910OP•7mo ago

it's really weird, the code now also fails with a normal Docker image on another VPS. So I don't think the issue is particularly related to Dokploy. I created an issue with a minimal sample at the Chrome wrapper lib that I'm using https://github.com/YusukeIwaki/puppeteer-ruby/issues/339

GitHub

Help needed: Pupeteer.launch hangs inside Docker image · Issue #33...

I have an issue getting a headless Chrome browser to run in my Ruby on Rails app deployed via Docker. It times out on all requests when started from the web server, but it works perfectly when laun...

Gaming

Programming

Headless Chrome Timing Out on Dokploy Deployment

Did you find this page helpful?