网关超时在Apache上使用Django和mod_wsgi

Question

I'm having a problem getting Apache to reply properly. I'm getting a browser error: Gateway Timeout: The gateway did not receive a timely response from the upstream server or application.

I know the application is connecting through mod_wsgi to the Django layer, as I've verified queries are occurring in PostgreSQL when the connection occurs, from the Wagtail CMS's middleware running.

The system is running RHEL 7.1, Apache 2.4.6, and mod_wsgi 4.4.21 with Python 3.4 and Django 1.8.

Here's the error in the Apache log:

[Thu Jan 07 13:25:54.554395 2016] [wsgi:error] [pid 17128] [client 128.91.91.123:50664] Timeout when reading response headers from daemon process 'my_classroom-https': /var/www/html/my_classroom/my_classroom/wsgi.py

Here's wsgi.py:

import os, sys
from socket import gethostname
from django.core.wsgi import get_wsgi_application

sys.path.append(os.sep.join(os.path.abspath(__file__).split(os.sep)[:-2]))

# cpl == Core, Production, Linux..
if(gethostname()[:3]) == 'cpl':
    os.environ.setdefault("DJANGO_SETTINGS_MODULE", "my_classroom.settings.prod")
# csl == Core, Staging, Linux.
elif(gethostname()[:3]) == 'csl':
    os.environ.setdefault("DJANGO_SETTINGS_MODULE", "my_classroom.settings.stage")
# else use dev settings.
# cdl == Core, Development, Linux. vag == Vagrant.
else:
    os.environ.setdefault("DJANGO_SETTINGS_MODULE", "my_classroom.settings.dev")

application = get_wsgi_application()

And here's the relevant portion of the Apache config:

LoadModule wsgi_module modules/mod_wsgi.so
LoadModule ssl_module modules/mod_ssl.so

WSGISocketPrefix /var/run/wsgi

NameVirtualHost *:443
Listen 443
<VirtualHost *:443>

  ServerName my-python-dev.school.edu
  ErrorLog /var/www/logs/classroom-apache-errors.log

  SSLENGINE on

  SSLCertificateFile /etc/pki/tls/certs/localhost.crt
  SSLCertificateKeyFile /etc/pki/tls/private/localhost.key
  SSLCipherSuite ALL:!ADH:!EXPORT:!SSLv2:RC4+RSA:+HIGH:+MEDIUM:+LOW
  SSLProtocol all -SSLv2

  WSGIApplicationGroup %{GLOBAL}
  WSGIDaemonProcess my_classroom-https python-home=/var/www/virtualenvs/my_classroom request-timeout=600
  WSGIScriptAlias /classroom /var/www/html/my_classroom/my_classroom/wsgi.py process-group=my_classroom-https
  WSGIProcessGroup my_classroom-https
  Alias /classroom/static/ /var/www/html/my_classroom/static/

  LogLevel info
</VirtualHost>

I've tried a few few things, including increasing the request-timeout parameter of the WSGIDaemonProcess directive to be 600 seconds instead of 60, but it still seems to timeout right around 60 seconds.

Any ideas or help would be greatly appreciated. Thanks.

Answer 1

I got this working. The problem was a secondary database definition (not default ) which couldn't connect to the server. There was no indication of this anywhere in the logs, unfortunately. This is a tough error to debug.

Here's some steps I'd take next time I see this error:

• Take a very deep breath and relax!
• Confirm if any activity is happening in the database at all for the request.
• Make sure you can connect to all services needed by telnet ( telnet mypostgresserver.domain.com 5432 , telnet ldap.myprivate.network 636 ), and that they're not hanging.
• Modify the settings files, excluding anything possible (middleware, optional apps, backends other than ModelBackend, etc)
• Failing that, start commenting out segments of the settings files, to see if you can get a less cryptic response.