1

I am trying to serve a Django app using gunicorn. In my current setup running on Debian 6 (production) and Ubuntu 11.04 (development), gunicorn (v0.12) is behind nginx and managed by supervisor. The setup runs OK, except that exceptions are not logged.

It is the same if I run gunicorn from the command line: manage.py run_gunicorn 127.0.0.1:XXXX. Again there's no output. The problem applies only to ERROR-level events, INFO-level output is produced when workers get started. The traceback is displayed on the page if DEBUG = True.

The app runs on Django v1.3 with no changes in the default logging settings.

Today v0.13 featuring a refactor of logging has become available, but the issue is still there.

I'd very much appreciate any help.

Edit: I got error emailing working, it is irrelevant to the logging issue.

shanyu
  • 111
  • 6

1 Answers1

0

Do you mean that you get no log message upon an error 500? After about 30 seconds into a request? In that case, gunicorn killed off its child process that was serving you. And it kills in a quick and hard way that leaves no logging.

Reinout van Rees
  • 170
  • 6