Cherrypy app on proxy port errors

treepple

I have a Cherrypy application on a proxy port developed to manage video resources and provide an RSS feed. It is used to upload videos to AWS S3 and store info on AWS RDS. This is an application that has been moved from Webfaction recently where it worked for more than one year without issues.

The application is logging errors (in the site logs) I can't seem to track down. The errors do not seem to be logged on all pages, but only a few - the RSS feed being one of them. They are though consistent.

This is the error:

[error] 31781#31781: *8133 connect() failed (111: Connection refused) while connecting to upstream

As far as I can tell and after a decent amount of searches it looks like it has something to do with how Nginx is set up, but can't figure it out. Does anyone have any idea how to approach this? Or even if I can do it on my own?

Thanks!

Gelu

sean

treepple If you've not done so already, please try server.socket_host = '127.0.0.1' in your CherryPy config and then restart the application.

treepple

sean Thank you for responding this fast! The host is currently set to '127.0.0.1'. I have other apps that are set to '0.0.0.0' that do not seem (as far as I can tell) experience the same issues. I'll try that and see where it goes.

treepple

sean Unfortunately the change to '0.0.0.0' did not make a difference. I am pursuing other possibilities, but if you have any guidance I would appreciate the help. I may have to recurse to opening a ticket when I exhaust all the remedies (I know of) on my end.

What port is Nginx set up to respond to?

treepple

sean There is definitely something going on. I checked multiple applications I have running on Cherrypy on two different vps servers I manage on Opalstack and they all seem to experience the same behavior. The websites seem to work fine in the end, but I would still like to resolve these errors reported in the site logs.

I enabled development mode on one of the Cherrypy applications and the log does not show any errors being reported. Therefore, I assume the Cherrypy application itself works fine. The fact that the errors are reported in the site logs seems to indicate (as far as I can understand your setup) that they happen because of Nginx or something related to it.

Thank you in advance for any support that you can provide in resolving this issue.

sean

treepple What port is Nginx set up to respond to?

Our front-end Nginx listens on 80/443 like a normal web server. When it receives a request for a domain that's attached to one of your sites, it proxies it to whatever app you have assigned via your Opalstack dashboard.

I think what might be happening is that somehow requests are going to your app over IPv6, which we're not fully set up with yet. Opening a ticket sounds like a good move, when you do let us know the shell user and app name so we can see what's going on.

treepple

sean Thank you! Opened the ticket. Hope you can help me find a solution.