Node.js app fails to start (Passenger) – dead-run.helioho.st – need error details from server logs

[smel]

 

 

Hello,

I am hosting a Node.js API on dead-run.helioho.st (HelioHost / Plesk). The app never starts: every request shows the generic Phusion Passenger error page (“We’re sorry, but something went wrong” / “Web application could not be started by the Phusion Passenger application server”). The Error ID changes on every refresh, so I assume the process crashes on startup each time.

My setup

• Application root: /httpdocs
• Document root: /httpdocs
• Application startup file: server.js
• Node.js version (panel): 23.11.1
• Package manager: npm (I run NPM install from the panel and Restart App)
• Custom environment variable: LEADERBOARD_JSON_ONLY=true (to avoid native SQLite; the app can fall back to a JSON file)

What I have already tried

• Uploaded a CommonJS server.js with module.exports = app (no listen() under Passenger).
• Dependencies: Express 4 and cors in package.json; removed native better-sqlite3 from the critical path when using the env var above.
• Deleted node_modules and ran a fresh npm install, then Restart App.
• Confirmed server.js, package.json, and node_modules are under /httpdocs.
• Cleared browser cache; waited after changes.

What I need from you

1. Please check the Passenger / Node.js logs for my account/domain and tell me the real stack trace (e.g. missing module, permission denied on data/, Node version issue, etc.). I do not have access to /var/log/passenger/passenger.log from my side.
2. If something in my Plesk Node.js settings is wrong (application root vs document root, Node version, or permissions on /httpdocs / data/), please advise what to change.

Sample Error IDs (if helpful): I can provide the latest ID from the Passenger page if you tell me where you prefer I paste it.

Thank you for your help.

Best regards,
smel
Domain: dead-run.helioho.st

 

[Krydos]

Do you have your environment set to development or productions?

[smel]

where do i see my environment ?
i founded
its production

do i need to change to development ?

Edited 13 hours ago by smel

[smel]

is this it

Edited 13 hours ago by smel

[KazVee]

58 minutes ago, smel said:

do i need to change to development ?

The error message you posted says:

"You can also get a detailed report to appear directly on this page, but for security reasons it is only provided if Phusion Passenger is run with environment set to development and/or with the friendly error pages option set to on."

So if you want a detailed report to appear, then you need to set the environment to development.

Since this is the second time you've been impatient about the unpaid volunteer staff not being able to help you immediately, I'd recommend you make use of the technical documentation in our Wiki: https://wiki.helionet.org

There is a search function and detailed information on how to use Node on the shared hosting servers. 

It's unreasonable to expect any of us to drop everything at our paying jobs or school classes because you're not patient enough to wait for someone to be available to assist you, for free, in our spare time. 

[smel]

ok thx i will try to change to development to see if i can see the error

both https://dead-run.helioho.st/api/scores#

and https://dead-run.helioho.st/api/health#

are getting 504 Gateway Time-out sometimes and when it loads its still the in the same state even that i already change to development

it says i dont have permission to acess the log but i already switch to development

Edited 11 hours ago by smel

[wolstech]

It takes up to 2 hours for a change to development mode to take effect. Wait 2 hours then try again and it should provide more information.

If for some reason it does not, we can escalate it at that point so Krydos can take a look at the logs for you.

[smel]

thx so i will check tomorrow again to try to see the logs

[smel]

Still doesnt show the error and now its says 504 Gateway Time-out doesnt load the page

https://dead-run.helioho.st/api/health#

https://dead-run.helioho.st/log/development.log

and the log page stills says

i cant see the logs yet if  Krydos can take a look at the logs for me i would apreciate very mutch

Update now the page is showing the error not api health but api scores

I now have detailed diagnostics: Passenger times out at “Handshake with subprocess — 90s”. My Node process prints a log line but never reaches “app start listening for requests”. 
I realized my app was not calling listen() under Passenger. I’ve updated it to start the HTTP server with server.listen('passenger') when IN_PASSENGER=1 / Passenger env vars are set.
I’ll redeploy and restart; this should fix the spawn timeout.

Hi! I fixed it — the issue was Passenger spawn timeout because the app wasn’t starting to listen under Passenger. I updated the Node app to listen via the Passenger socket (server.listen('passenger') when IN_PASSENGER=1 / Passenger env vars are present).
After redeploy + restart, /api/health returns JSON successfully (ok: true). Thank you for the guidance.

thx 4 the help

Edited 1 hour ago by smel