Replit failing promotion

Deployments were working fine for a while, then all of the sudden got this error

link (private): https://replit.com/@hearty/Pluto

tried “poetry update” “pip install psycopg2” “pip install psycopg2-binary” and “poetry add psycopg2-binary” per some suggestions on similar issues, but still getting the same error when trying to deploy. works fine when i click run.

code snippet
1 Like

I am actually getting the same error. It seems to only happen in python, I have tried flask, django, and fastAPI for web apps and the same error.

I did some digging. I couldn’t solve the issue, but it looks like pid1 is a binary layer and hosting is a layer aswell pid1 stands for process id 1. It seems like it is a issue with replit.

I am getting same error over 48hrs now

I think my error with “mastodon” is quite the same … (another bug report and no issue during the deployment 2 days ago)
I would say that deployment doesn’t take all the modules from the dev repl

@hearty @rexhall @GRANDS-ROISROIS @fvillemin The issue is most likely with our package caching layer which was recently rolled out to explorers. There are a few bugs that have been exposed here that we’re going to fix before rolling this out to everyone. I would recommend taking yourself out of the explorer role in order to deploy. You can do this by going to your Replit account page, scrolling down to the “roles” section and uncheck the “Explorers” box. Let me know if it still doesn’t work after you do that.

Hello @lincoln-replit , i’m not Explorer anymore for some time (too much issue with new features) so it should not be that reason for me … :frowning:

I see your other post, I will respond there.

This is how mine look, i have not change it yet, do i need to do anything

It looks like we also had it rolled out to a percentage of users in addition to just explorers. We’ve rolled it back now. If you reload the page and try again it should work as before.

The issue persists even though I am not an explorer.

I deployed a few python repls doing various things through auto scale. None worked, they all had the same error, even though some didn’t even have a web server.
I then used this code:

import socket

def print_open_ports():
		for port in range(1, 65535):
				sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
				sock.settimeout(1)
				result = sock.connect_ex(('localhost', port))
				if result == 0:
						print("Port {} is open".format(port))
				else:
					  print("X")
				sock.close()

print_open_ports()

And it printed all ‘X’ meaning there were actually no open ports, I may have coded this incorrectly, but this doesn’t seem right. I have tried with explorer on and off.

Hi @rexhall, the easiest thing will probably be to schedule a debugging session here: Calendly - Replit Deployments Team

Ok, thank you. I will schedule one at some point.