Python module being reloaded for each request with django and mod_wsgi
I have a variable in init of a module which get loaded from the database and takes about 15 seconds.
For django development server everything is working fine but looks like with apache2 and mod_wsgi the module is loaded with every request (taking 15 seconds).
Any idea about this behavior?
Update: I have enabled daemon mode in mod wsgi, looks like its not reloading the modules now! needs more testing and I will update.
Answer
It seems like you're encountering an issue with the module being reloaded on every request when using Apache2 with mod_wsgi, which results in the 15-second delay due to the database query in the module's __init__
method.
Why This Happens:
When using Apache2 with mod_wsgi, there are two main modes for handling the application:
-
Daemon mode: This mode runs your Django application in a separate daemon process, and the application is not reloaded for every request. This is the mode you want to use for performance.
-
Embedded mode (default): In this mode, mod_wsgi directly embeds your Django application inside the Apache worker processes. This can lead to the module being reloaded on each request, especially if you import it inside a module or during the initialization phase.
In embedded mode, Django (and by extension, the Python module containing your variable) may be loaded multiple times, which causes the initialization code (like the one where you query the database) to run repeatedly for each request.
The Solution:
You seem to have already enabled daemon mode in mod_wsgi, which should reduce the overhead and prevent reloading the module on each request. However, it requires proper configuration to ensure that the module is loaded only once and shared across all requests.
Key Considerations for Daemon Mode:
-
Ensure daemon mode is properly configured: Make sure that you've configured mod_wsgi with daemon mode correctly. This can be done by adding
WSGIScriptAlias
andWSGIDaemonProcess
in your Apache configuration.Example configuration:
WSGIDaemonProcess your_project python-path=/path/to/your/project python-home=/path/to/virtualenv WSGIProcessGroup your_project WSGIScriptAlias / /path/to/your/project/your_project/wsgi.py <Directory /path/to/your/project> Require all granted </Directory>
-
Use
mod_wsgi
for process sharing: Daemon mode uses a pool of processes to serve requests. This means that the same Python process (and its modules) should be reused across multiple requests. -
Check if
__init__
is running on each request: Even with daemon mode enabled, it's important to ensure that the database query (or any heavy operation) is not being executed on each request. Instead, you may want to cache the result of that query for the lifetime of the process. Consider storing the result in a global variable, or using an external caching mechanism (e.g., Redis or Django's caching framework).
Caching the Data:
If the data from the database only needs to be loaded once per process or session, you can cache it using one of the following approaches:
-
Django Cache Framework: Use Django's built-in cache framework to cache the database result at the module level.
Example:
from django.core.cache import cache class MyModule: def __init__(self): # Check if the data is already cached self.data = cache.get('my_data_key') if not self.data: # If not cached, fetch from DB self.data = self.fetch_data_from_db() # Cache the data for 1 hour (or as needed) cache.set('my_data_key', self.data, timeout=3600) def fetch_data_from_db(self): # Your 15-second DB query here return some_database_query_result
-
Global Variable or Singleton Pattern: You can use a global variable or a singleton pattern to store the data and ensure it is loaded only once per process.
Example:
class MyModule: _data = None def __init__(self): if MyModule._data is None: MyModule._data = self.fetch_data_from_db() self.data = MyModule._data def fetch_data_from_db(self): # Your 15-second DB query here return some_database_query_result
Debugging Further:
If you still experience issues after switching to daemon mode, here are a few debugging steps:
- Check the Apache logs: Look for any errors or warnings in the Apache logs (
/var/log/apache2/error.log
) related to mod_wsgi or the application. This might give you clues about misconfiguration. - Test module loading: You can add a simple print statement or logging at the top of your module to see how often it is being reloaded. This can help you confirm if the issue persists even with daemon mode enabled.
- Check your
mod_wsgi
configuration: Ensure that all the necessary mod_wsgi settings for daemon mode are correctly configured, particularly theWSGIDaemonProcess
andWSGIProcessGroup
.
Summary:
- Make sure daemon mode is enabled and correctly configured.
- Use caching to avoid running the database query on every request.
- Ensure your database query only runs once per process (using global variables or the Django cache framework).
By applying these methods, you should be able to significantly reduce the loading time and avoid re-running expensive operations on every request.