When you think of reverse proxies, you say nginx. nginx is the ideal reverse proxy tool.
But now the conditions are tough. The server doesn't have nginx and doesn't have root privileges, which means you can't compile and install nginx, and only one port, 80, is open for access. How do you get a request on port 80 to be forwarded to an http service on another port?
In other words, without nginx, how do you get a request from http://localhost:80/new_req to be forwarded to the http://localhost:9999 service, just as if you were accessing http://localhost:9999 directly?
Well, that requires implementing our own reverse proxy functionality.
This article shares how to implement a reverse proxy using only Django.
1.Installation
pip install django-revproxy
This wheel has a corresponding whl file, which is very easy to install successfully, and does not depend on other wheels. If it is an intranet environment, download the whl file from pypi.org, copy it into pip and install it.
2.Configuration
In Django's configuration file settings.py, in INSTALLED_APPS, add 'revproxy'.
#Add 'revproxy' to INSTALLED_APPS.
INSTALLED_APPS = (
# ...
'django.contrib.auth',
'revproxy',
# ...
)
Then write a view class, such as myproxy/views.py:
from revproxy.views import ProxyView
class TestProxyView(ProxyView):
upstream = "http://localhost:9999" ##http://example.com
Then modify Django's urls.py file to add a route like this
from myproxy.views import TestProxyView
from django.conf.urls import url
urls = [
# ...
url(r'^new_req/(?P<path>.*)$', TestProxyView.as_view()),
# ...
]
Or, instead of creating a new views.py file, you can just write this in url.py
from django.urls import re_path
from revproxy.views import ProxyView
urlpatterns = [
re_path(r'(?P<path>.*)', ProxyView.as_view(upstream='http://example.com/')),
]
Finally, deploy the Django service on port 80, and then visit
http://localhost/new_req
is equivalent to accessing
http://localhost:9999
3.How it works
This diagram is what django-revproxy does：
1.Django receives requests from clients and processes them using revproxy.proxy.
2.Revproxy will clone client requests.
3.If the user is authenticated in Django and the add_remote_user property is set to True, the HTTP header REMOTE_USER will be set to request.user.username.
class CustomProxyView(ProxyView):
upstream = 'http://www.example.com'
add_remote_user = True
4.If the add_x_forwarded attribute is set to True, the HTTP headers X-Forwarded-For and X-Forwarded-Proto will be set to the IP address and protocol (http or https) of the requester, respectively.
5.The cloned request is sent to the upstream server, which is upstream.
6.After receiving a response from upstream, the view will process it to ensure that all headers are set correctly. some headers, like Location, are considered special cases.
7.The response received from the upstream server is converted to django.http.HttpResponse. for binary files, use StreamingHttpResponse to reduce memory usage.
8.If the user sets a set of diazo rules and a theme template, the diazo/XSLT transformation is applied to the response body.
views.py :
from django.urls import re_path
from revproxy.views import DiazoProxyView
proxy_view = DiazoProxyView.as_view(
upstream='http://example.com/',
html5=True,
diazo_theme_template='base.html',
)
urlpatterns = [
re_path(r'(?P<path>.*)', proxy_view),
]
base.html :
<html>
<head>...</head>
<body>
...
<div id="content"></div>
...Fix all links in the docs (and README file etc) from old to new repo
9.Finally, the response will be returned to the user.
4.django-revproxy
To share a file, I use a Python command in a directory: python -m http.server, which allows files or folders in that directory to be shared via http for others to download and use, and django-revproxy, which allows this functionality to be integrated into one of Django's routes in The developer server deployment is really easy for everyone to use.