I have tried following guides like this one but it just didnt work for me.
So my question is this: What is a good guide for deploying Django, and how do you deploy your Django.
I keep hearing that capastrano is pretty nifty to use, but i have no idea as to how to work it or what it does (apart from automation of deploying code), or even if i want/need to use it or not.
mod_wsgi in combination with a virtualenv for all the dependencies, a mercurial checkout into the virtualenv and a fabric recipe to check out the changes on the server.
I wrote an article about my usual workflow: Deploying Python Web Applications. Hope that helps.
I have had success with mod_wsgi
In my previous work we had real genius guy on deployment duties, he deployed application (Python, SQL, Perl and Java code) as set of deb files built for Ubuntu. Unfortunately now, I have no such support. We are deploying apps manually to virtualenv-ed environments with separate nginx configs for FastCGI. We use paver to deploy to remote servers. It's painful, but it works.
This looks like a good place to start: http://www.unessa.net/en/hoyci/2007/06/using-capistrano-deploy-django-apps/
I use mod_python, and have every site in a git repository with the following subdirs:
mysite
template
media
I have mysite/settings.py in .gitignore, and work like this:
do development on my local machine
create remote repository on webserver
push my changes to webserver repo
set up apache vhost config file, tweak live server settings.py
run git checkout && git reset --hard && sudo /etc/init.d/apache2 restart on webserver repo to get up-to-date version to its working copy and restart apache
repeat steps 1, 3 and 5 whenever change request comes
The easiest way would be to use one of the sites on http://djangofriendly.com/hosts/ that will provide the hosting and set up for you, but even if you're wanting to roll your own it will allow you to see what set up other sites are using.
Related
Comming from PHP(without frameworks), I don't quite understand how deploying works in Python. I have completed my simple Django version: (1, 10, 1, 'final', 1) blog project and all I have to do is put it online. I am using linux openSUSE distro and virtualenv
I have access to a mysql database with phpmyadmin and I have some space, accessed with filezilla. hosting site: https://host.bg/
But then I started researching of how to deploy my project and I stumbled upon stuf like apache, Nginx, wsgi and other stuff I haven't used in the project and not quite familiar how they work.
So my question is: Can I make my project into a folder with some files in it, copy->paste them in filezilla and have a working site and if not, how does django deployment really work and what should I do from here ?!
I would recommend you to use Git instead of FTP protocol. As you are using Linux you can easily connect to your SO using ssh.
About the deployment, I would recommend you to use GUnicorn for a WSGI way.
It's not hard to deploy with, but if you get in trouble you can use the official Django documentation for deploying Django with WSGI:
Link
Ps.: As you are using Linux, I would recommend to you to use VirtualEnv to allow you server many Django sites in the same Linux instance with the isolated environments
Check what version of Python is installed on the server hosting your account and if there's option for ssh access.
Host.bg and Bulgarian hosting providers in general fail to keep up with most things other than php and mysql. With shared plans they avoid installing anything new too.
I'd say contact support and see if they would be able to assist you with Apache configurations and whatever else is needed for your project.
I'm wondering what steps are evolved in moving a django project to a new server.
basicly i'm completely new to Django and have a few questions. The server it is on is now is not stable so I need to act fast. I did not build the app that is there but have pulled down the www folder from the root server. The server is running centOS.
Questions.
is Django backwards compatible or will I need to insure that the same version is installed?
Apart from moving the files what other steps are involved in running the app?
Will I need to use centOS or will any linux server do?
I have a database cluster of PostgreSQL ready to go also.
Start with the docs here - this will give you a good overview.
To your specific questions:
1/ Django is not backwards compatible. You should install 1.6.x. Likely, there's a requirements.txt file in the root directory of your app. On your new server, install pip and then pip install -r requirements.txt will install your dependencies. I would personally use virtualenvwrapper to manage your dependencies on the server
2/ Check the docs, but the main steps are:
Choose a web server. I personally use nginx. You'll need to setup your nginx.conf.
Choose a Python WSGI HTTP Server - I use gunicorn. You'll also need to configure this. This tutorial is a great place to start.
If you use the DigitalOcean tutorial above, any linux server will do. Last, you'll need to upload your Postgres database to the server but sounds like you're able to do that.
3/ You will need to edit your settings.py of the Django project and update certain variables.
If you're changing your database, as well as the app deployment, you'll need to edit the database connection, run ./manage.py syncdb and ./manage.py migrate (if you're using South) to set up the database schema.
It's also recommended to change the SECRET_KEY between deployments.
If you're deploying on a different hosts, you'll need to edit ALLOWED_HOSTS appropriately for your new deployment as well.
Good luck!
I have a Django project with 1 app that is working locally and I am trying to make it work on the server but I imagine I am still missing something...
The steps I have followed are:
1) create a virtualnev
2) Install django and the libraries I need
3) copy my local project to the server, keeping the same directory structure
4) create the file passenger_wsgi.py (python passenger_wsgi.py did not return any error)
After this do I need to do anything like python manage.py runserver? Or with this I should be already able to see the site through mydomain/my project/ my app (when I do ot just get an error 404)?
I have read the django book and followed the tutorial, but this part is not well described anywhere...
Thanks in advance for any help!
Deployment is explained in the documentation.
You need to actually serve your application with some kind of HTTP server and something to run your python code. Some of the possible combinations are:
nginx with uWSGI
Any web server as reverse proxy with gunicorn
Apache with mod_wsgi
Your hosting service may or may not give you the choice or even the possibilty to do this.
There is a list of Django friendly hosters in the Django wiki.
Well firstly you might want to go through the previously asked questions.
When you are deploying using passenger you do not need to run manage.py runserver , etc. The passenger_wsgi file should take care of it.
You might want to check this link out, in the first answer also contains link to Dreamhost which details quite extensively on how to achieve the same.
Visit Deploying Django app using passenger
From my personal experience I found Nginx and uwsgi setup to be much more easier to handle and work and debug in the logs, but this is subjective to your needs and platform you may have.
Expanding on this question I am trying to deploy Django on OpenShift but I'm having some problems understanding OpenShift.
I have managed to get as far as setting up a quick app with the git repo https://github.com/openshift/django-example but have the following questions:
Can I develop locally after git cloning to my local? (virtualenv, adding packages)
Packages, what's the deal? local, remote, adding, sycing, virtualenv, git, ...
I came across this line in Nate Aune's PaaS Bakeoff (slide 42) for setup.py and it looks quite useful:
install_requires=open('%s/project.txt' % \
os.environ.get('OPENSHIFT_REPO_DIR', PROJECT_ROOT)).readlines(),
(because I know I can pip freeze > requirments.txt in my virtualenv)
... Is %s/project.txt in wsgi or tthe directory below wsgi? Do I have to set PROJECT_ROOT with some funky os stuff?
EDIT
Basically:
Is it best to ssh into your OpenShift application (lest say you have a dev one) and work there or work off a local copy?
How do you install python packages after you have ssh'ed in to your OpenShift app? (virtualenv)
If you ssh'ed into your OpenShift app do you have to do anything after: creating a project, creating an app (manage.py startapp ...), changing code in your django app
If local is the best option:
How do I use the example locally?
Do I need to setup a virtualenv to work locally?
How do I make sure the python packages django needs are on OpenShift?
How do I add python packages to my OpenShift version (I'm presuming git doesn't do that)
I would suggest developing in a local copy.
In my experience there is no reason to ssh in, unless if you
need to perform one-of operation (migrating databases for example)
The requirements for the openshift app are specified in the setup.py file.
install_requires=['Django>=1.6.0', 'redis>=2.0']
Locally you can of course work in a virtualenv
I was facing a similar situation and found the answer provided by Luis Masuelli to the question posted here very informative. Hope it might help you.
I'm creating a server with Apache2 + mod_python + Django for development and would like to know how to use Mercurial to manage application development.
My idea is to make the folder where the Mercurial stores the project be the same folder to deploy Django.
Thank you for your attention!
I thought about this, good idea for development.
Use mercurial in common way. And of course you need deploy mercurial server before.
If you update your django project, it will be compiled on the fly.
My workflow:
Set up mercurial server or use bitbucket
Init repo locally
Push repo to central repo
On server pull repo in some target dir
Edit smth locally and push to central repo
Pull repo on server and everything is fine