This should not last so long, even if you used the suboptimal configuration using the built-in mode and Apache prefork MPM. Although, you could do something worse if you have MaxRequestsPerChild set to 1 for Apache processes.
The following recommendations are suggested: make sure that you use the mod_wsgi daemon mode and start using a single multi-threaded process (the default settings for WSGIDaemonProcess). This will ensure at least that there is only one instance of Django and not potentially one per Apache server process.
As a confirmation of what you are doing, edit your question and post a snippet of the current Apache configuration that shows how to configure the mod_wsgi bits. Also include whether you are using the Apache pre-program or the working MPM that is detected when 'httpd -V' is run and which platform you are using.
By the way, have you tried the simple WSGI world greeting program to test your installation? See ' http://code.google.com/p/modwsgi/wiki/QuickConfigurationGuide '. And have you tried with an empty Django site, not your real one, to find out if your changes are?
source share