I'm having some problems trying to deploy a new Django application on Webfaction. The Django1.8-Python3.4 option has disappeared (I already have an app deployed under this configuration on the same server) and Django1.10-Python3.5 shows some rare messages in the error log like "TEMPLATE_LOADER attribute not specified in settings" when trying to browse my website
When I restart apache this is the first two messages in the log:
[Sun Feb 05 17:58:45.210618 2017] [mpm_worker:notice] Apache/2.4.23 (Unix) mod_wsgi/4.5.9 Python/3.5 configured -- resuming normal operations [Sun Feb 05 17:58:45.210644 2017] [core:notice] Command line: '/home/***/webapps//apache2/bin/httpd.worker -f /home//webapps/***/apache2/conf/httpd.conf'
(Where **** are my user and project names. )
I've configured everything to use a virtualenv with Python3.4 and everything goes well under interactive shell (sync database, etc..). On the other hand I can reproduce the errors appearing in the logs in a interactive shell under a virtualenv with Python3.5 and Django 1.10.
What does Webfaction configure behind the scenes when I create a Python3.5 project. Why does appear Python3.5 as default Python (Python/3.5 configured)? Why my log shows errors corresponding to Python 3.5/Django 1.10 when everything is pointing to my virtualenv? (some hours spent checking everything is ok, libs etc..)
A lot of things happen behind the scenes in the installation scripts. The basic difference between the default Django app and a generic mod_wsgi/python app is the extra configuration for Django. The application binaries and libraries are the same. If your code is only set to run at Django 1.8 than you can try to Downgrade by following the upgrade procedure but using a lesser version. You may also want to check the Django version release notes to see if you can fix your code to work on 1.10
The line from the log you are posting is only the reboot resume, it is not enough of the message to debug.
answered 07 Feb, 22:17