I have a configured and running setup that I am looking to optimize. I do not want to swap out Apache for gunicorn or other options at this stage.
My setup is so:
Ubuntu 11.04
Default nginx from apt-get
Default apache from apt-get
Nginx serves static files, and passes application requests through to Apache. Apache will have between 5-8 Django projects (ie - distinct websites). Small to medium traffic. Apache only has django projects (served via mod_wsgi) - I don't need php or anything that Django does not need.
From the default Ubuntu/Apache, what mods can I disable, and are there any other configuration tweaks I can do to more optimally use resources on my machine.
One configuration tweak is that, if you don't need apache to serve anything else than django sites, ditch apache completely.
Use a dedicated WSGI server like uwsgi ( http://projects.unbit.it/uwsgi/ ) or gunicorn ( http://gunicorn.org/ ). They are well documented and so low on resource usage.
You can use WSGIOptimize option to make all your .py files into .pyo. You can also use Memcached to enable cache. This blog describes how to run several django sites in one daemon. It is said to be useful in case where you need to serve a lot of sites, each with low traffic. This contains tips from Jacob Kaplan-Moss.
Related
The question is in the title.
I need to deploy a Django application in a local network (i still don't know how to do but i suppose it's quite easy) but i still need to develop it. My question is how to do to allow users to use the application while i'm still developing it ?
Is it a solution to keep to versions of the application, one deployed and one in development ? In this way, I can replace the application deployed by the newly developed one when I finish coding it.
Another question concerns the database, can I still modify the database if I just add new models without touching existing ones ?
Thank you in advance,
This is a good blog entry that covers deploying Django using Heroku. I'll give you a quick rundown of what makes all the different technologies important:
Git
Git, or any other version control system, is certainly not required. Why it lends itself to deploying Django projects is that your usually distributing your application by source, i.e. your not compiling it or packaging it as an egg. Usually you'll organize your Git repository as such that updating your application on the server only requires you to do a checkout of the latest sources--nothing else.
virtualenv and pip
This, again, is not a strict requirement, but I'd strongly suggest you take the time to familiarize yourself with virtualenv and pip if you already haven't done so, since it's going to make deploying your Python applications across different runtime environments, local or remote, a breeze.
Basically, your project will need to have at least Django and Gunicorn available on the Python path, possibly even a database driver. What that means is that every time you try to deploy your application somewhere you'll have to install Python and do the easy_install dance all over.
virtualenv will redistribute a Python installation, which in turn means that the new Python instance will, by default, have it's very own Python path configuration relative to the installation. pip is like easy_install on steroids, since it supports checking out Python dependencies directly from code repositories and supports a requirements file format with which you can install and configure all of your dependencies in one fell swoop.
With virtualenv and pip, all you'd need to do is have a simple text file with all your dependencies that can be parsed with pip and an installed Python distribution on the machine. From there you just do git checkout repo /app/path; easy_install virtualenv; virtualenv /app/path; ./app/path/scripts/activate; pip install -r /app/path/requirements.txt. Voila, Gunicorn, Django and all other dependencies are then installed and available immediately. When you run the Gunicorn Django script with the Python instance in /app/path/scripts, then the script will immediately have access to the Gunicorn sources and it will be able to locate your Django project which will have access to Django and other dependencies as well.
Gunicorn
This is the actual Python application that will manage your Django instance and provide an HTTP interface that exposes it to HTTP clients. It'll start several worker processes which will all be distinct Python virtual machines loaded with the sources of your application and it's dependencies. The main Gunicorn process will in turn take charge of managing which worker processes manage which requests for maximum throughput.
The basic principle of wiring Nginx and Gunicorn
The most important thing to observe is that Nginx and Gunicorn are separate processes that you manage independently.
The Nginx Web server will be publicly exposed, i.e. it will be directly accessible over the internet. For requests to static media, such as actual images, CSS stylesheets, JavaScript sources and PDF files accessible via the filesystem, Nginx will take charge of returning them in the response body to HTTP clients if you configure it to look for files on the path where you configured your project to collect static media.
Any other request should be proxied to your Gunicorn instance. It will be configured to listen to HTTP requests on a certain port on the loopback interface, so you'll using Nginx as a revers proxy to http://127.0.0.1:8080 for requests to your Django instance.
This is the basic rundown for deploying your Django projects into production that should satisfy the needs of 95% Django projects running out there. While I did reference Nginx and Gunicorn, it's the usual approach when it comes to setting up any Web server to act as a reverse-proxy to a Python WSGI server.
I have a django application running on a server. I want to use let's encrypt to provide an encrypted connection. I could use the standalone option of their ACME client, but i don't want to stop my server, what i would have to do.
So there is the webroot option, that work with my allready running webserver (nginx). Django would process the request in this case. My question is, how should it look like on the django side to get this running (keeping automated renewal several months in mind)?
I don't know what setup others use, but I generally set up Django apps with Nginx serving static content and Gunicorn as the application server. It's widely accepted that Django apps usually use this kind of two web server setup. The standard instructions for setting up Let's Encrypt with Nginx worked fine for me.
Or Digital Ocean have an excellent guide too.
EDIT: It looks like Nginx can do a "graceful" reload that just updates the config with no downtime. For Debian or Ubuntu pre Systemd this would be sudo service nginx reload, while for a distro with Systemd the command is sudo systemctl reload nginx.service.
In case other users come this way like I did from Google, here's how I improved this situation:
I was unsatisfied by my options when it came to creating ACME challenges for Let's Encrypt when running a Django application. So, I rolled my own solution and created a Django app! Basically, you can manage your ACME challenges as just another object, and the app will produce the proper end-point URL.
Yes you are installing an app which means a deploy / update to your app, but once you've done that managing your challenges is far easier in the long run.
Simply pip install django-letsencrypt and follow the README to be on your way.
I want to deploy a django site (it is the open source edx code on github).
I am faced with choosing between using
Apache with mod_wsgi
nginx with gunicorn
I have used Apache with mod_wsgi and it's cool enough, but i have no experience with the second option.
Which of these would be a better option in terms of speed and also to some extent, ease of use?
NB: I would need to run two different django sites on say, port 80 and 81 and access them from two different subdomains.
Nginx is a really light and easy to use solution and along with gunicorn it allows us to run any wsgi application and scale it easily.
Nginx is better at handling requests since it does not spawn a new process for every request unlike Apache.
I have written an answer on how to deploy django with nginx for a related question:
Deploying Django project with Gunicorn and nginx
Well,the few milliseconds you get with Nginx will not make a hudge difference regarding the time other processes take. Nginx may save RAM but it would only be a great difference on servers with a few RAM. For specific uses on big website there could be some more notable differences but this will become an expert affair then.
The real difference for most is probably the ease of learning. I don't find Apache to be specifically hard to use and the doc is clean. However most of Python tutorials I found are about using Nginx with Gunicorn.
If you already know how to use Apache with Python it would probably be more straight to the point to use it, unless you want to learn Nginx too to improve your CV.
However, if you are a newcomer, there is more documentation about Nginx with Python. It makes it the easier option.
I have good experience with nginx and gunicorn. They keep on working great when I've finally set all the settings right and got it running.
For nginx and gunicorn they are:
* nginx configuration files (/etc/nginx/sites-enabled/ and /etc/nginx/nginx.conf)
* gunicorn configuration files (/etc/init/gunicorn.conf and /etc/gunicorn.d/)
I've seen a tutorial for apache + mod_wsgi and it seems so much simpler to set up.
I have primarily worked with nginx and gunicorn. I am currently working with apache + mod_wsgi. It is actually easy if your Python version is 2.7 because mod_wsgi when installed directly from the package manager will work normally. But if your code is in a different Python version. mod_wsgi has to be built from source with the same version. If you installed your Python also from source then the procedure to get the whole web application working is fairly difficult.
Nginx and gunicorn on the other hand do not have any version issues, since the proxypass param makes it easy to forward requests to gunicorn. All we need to ensure is that gunicorn is running with the same version of Python that your code is in.
I have recently switched to Django for a web app I'm developing and I followed the instructions at Heroku for getting a Django app running on Heroku. I have a virtual environment in which my app is developed and I use git for version control and to push to Heroku. The link above suggests that I intall gunicorn:
The examples above used the default HTTP server for Django. For
production apps, you may wish to use a more production-ready embedded
webserver, such as Tornado, gevent’s WSGI server, or Gunicorn.
They then walk the user through installing Gunicorn.
My question is: what problems might I run into if I skip this step and just stay with the default? What benefits will Gunicorn give me?
Gunicorn is production ready and really easy to use. I use it for my websites. You usually should run it via a reverse proxy like Nginx. I'm not sure what Heroku is using. You really should try it.
In my experience it's much easier to use and configure than apache & mod_wsgi, and the other similar setups.
edit/update:
As a summary of the comments below, Heroku already uses Nginx as a reverse proxy
Much better performance, and probably better security and stability, too. Django's development web server (which is used by Heroku by default) isn't really designed to serve production applications.
django's server, is a development server . It is light weigh and easy to use but should not be used in production because it is not production ready. it cannot handle many requests. This link offers a comparison between gunicorn, uwsgi and django's development server.
I am quite a noob when it comes to deploying a Django project. I'd like to know what are the various methods to deploy Django project and which one is the most preferred.
The Django documentation lists Apache/mod_wsgi, Apache/mod_python and FastCGI etc.
mod_python is deprecated now, one should use mod_wsgi instead.
Django with mod_wsgi is easy to setup, but:
you can only use one python version at a time [edit: you even can only use the python version mod_wsgi was compiled for]
[edit: seems if I'm wrong on mod_wsgi not supporting virtualenv: it does]
So for multiple sites (targeting different django/python versions) on a server mod_wsgi is not
the best solution.
FastCGI can be used with virtualenv, also with different python versions, as you run it with
./manage.py runfcgi …
and then configure your webserver to use this fcgi interface.
The new, hot stuff about django deployment seems to be gunicorn. It's a webserver that implements wsgi and is typically used as backend with a "big" webserver as proxy.
Deployment with gunicorn feels a lot like fcgi: you run a process doing the django processing stuff with manage.py, and a webserver as frontend to the world.
But gunicorn deployment has some advantages over fcgi:
speed - I didn't find the sources, but benchmarks say fcgi is not as fast as the f suggests
config files, for fcgi you must do all configuration on the commandline when executing the manage.py command. This comes unhandy when running multiple django instances via an init.d (unix-like OS' system service startup). It's always the same cmdline, with just different configuration files
gunicorn can drop privileges: no need to do this in your init.d script, and it's easy to switch to one user per django instance
gunicorn behaves more like a daemon: writing pidfile and logfile, forking to the background etc. makes again using it in an init.d script easier.
Thus, I would suggest to use the gunicorn solution, unless you have a single site on a single server with low traffic, than you could use the wsgi solution. But I think in the long run you're more happy with gunicorn.
If you have a django only webserver, I would suggest to use nginx as frontendproxy, as it's the best performing (again this is based on benchmarks I read in some blogposts - don't have the url anymore).
Personally I use apache as frontendproxy, as I need it for other sites hosted on the server.
A simple setup instruction for django deployment could be found here:
http://ericholscher.com/blog/2010/aug/16/lessons-learned-dash-easy-django-deployment/
My init.d script for gunicorn is located at github:
https://gist.github.com/753053
Unfortunately I did not yet blog about it, but an experienced sysadmin should be able to do the required setup.
Use the Nginx/Apache/mod-wsgi and you can't go wrong.
If you prefer a simple alternative, just use Apache.
There is a very good deployment document: http://lethain.com/entry/2009/feb/13/the-django-and-ubuntu-intrepid-almanac/
I myself have faced a lot of problems in deploying Django Projects and automating the deployment process. Apache and mod_wsgi were like curse for Django Deployment. There are several tools like Nginx, Gunicorn, SupervisorD and Fabric which are trending for Django deployment. At first I used/configured them individually without Deployment automation which took a lot of time(I had to maintain testing as well as production servers for my client and had to update them as soon as a new feature was tested and approved.) but then I stumbled upon django-fagungis, which totally automates my Django Deployment from cloning my project from bitbucket to deploying on my remote server (it uses Nginx, Gunicorn, SupervisorD, Fabtic and virtualenv and also installs all the dependencies on the fly), all with just three commands :) You can find more about it in my blog post here. Now I even don't have to get involved in this process(which used to take a lot of my time) and one of my junior developers runs those three commands of django-fagungis mentioned here on his local machine and we get a crisp new copy of our project deployed in minutes without any hassle:)