Django CSRF Verification failed for admin panel

Question

I started a fresh Django 1.11 project with one app, one model and one admin panel. Locally, everything works. When I deploy it to Amazon EC2 and try to log in to the admin panel, I get a 403 (CSRF verification failed. Request aborted.). I see this in my debug log:

[WARNING] 2017-05-21 11:23:52,142 csrf 14263 140377210439424 Forbidden (Referer checking failed - Referer is insecure while host is secure.): /admin/login/

I inspected with Chrome's network utility the request, and I noticed that in my Request Header I have:

Cookie:csrftoken=hFhzOJPMOhkNWWWfRtlMOEum9jXV8XXWnOtw3OwZm2En9JUqYRVq632xyZfwSpzU

while in my Form Data I have:

csrfmiddlewaretoken:RHNpPfOHhg42FZnXmn9PZgNm3bN40C41XQZm4kvUP1oCSMl8tLJthFlxsR5FK4GZ

Should these two be the same? In my understanding they do, but when I try the same in my local environment, I see they're also not the same, but there it is working fine and I get the same token back in the Response Header as was sent in the Request Header, so I assume they don't need to be exactly the same? Note: I do not have a secure connection (https) at the moment, but will work on that after this is fixed.

I already tried/checked the following:

• Setting CSRF_COOKIE_DOMAIN (https://stackoverflow.com/a/42115353/1469465)
• Clear all cookies in my browser (https://stackoverflow.com/a/29574563/1469465)
• The variable CSRF_COOKIE_SECURE is not set and hence False (https://stackoverflow.com/a/29574563/1469465)
• My favicon is loaded correctly (https://stackoverflow.com/a/42021886/1469465)
• I have a DNS pointing to my EC2 instance, and have that subdomain in ALLOWED_HOSTS

Other answers I found on SO mention that you need to do something in the form itself, but this is a form from the Django framework.

Additional information

My nginx configuration from /etc/nginx/nginx.conf:

user www-data;
worker_processes auto;
pid /run/nginx.pid;

events {
    worker_connections 768;
    # multi_accept on;
}

http {

    ##
    # Basic Settings
    ##

    sendfile on;
    tcp_nopush on;
    tcp_nodelay on;
    keepalive_timeout 65;
    types_hash_max_size 2048;
    # server_tokens off;

    # server_names_hash_bucket_size 64;
    # server_name_in_redirect off;

    include /etc/nginx/mime.types;
    default_type application/octet-stream;

    ##
    # SSL Settings
    ##

    ssl_protocols TLSv1 TLSv1.1 TLSv1.2; # Dropping SSLv3, ref: POODLE
    ssl_prefer_server_ciphers on;

    ##
    # Logging Settings
    ##

    access_log /var/log/nginx/access.log;
    error_log /var/log/nginx/error.log;

    ##
    # Gzip Settings
    ##

    gzip on;
    gzip_disable "msie6";

    # gzip_vary on;
    # gzip_proxied any;
    # gzip_comp_level 6;
    # gzip_buffers 16 8k;
    # gzip_http_version 1.1;
    # gzip_types text/plain text/css application/json application/javascript text/xml application/xml application/xml+rss text/javascript;

    ##
    # Virtual Host Configs
    ##

    include /etc/nginx/conf.d/*.conf;
    include /etc/nginx/sites-enabled/*;
}

My site specific configuration from /etc/nginx/sites-enabled/MyDjangoService:

upstream MyDjangoService_wsgi_server {
  # fail_timeout=0 means we always retry an upstream even if it failed
  # to return a good HTTP response (in case the Unicorn master nukes a
  # single worker for timing out).

  server unix:/webapps/MyDjangoService/run/gunicorn.sock fail_timeout=0;
}

server {
    listen      80;
    server_name MyDjangoService;

    client_max_body_size 4G;

    access_log /webapps/MyDjangoService/logs/nginx_access.log;
    error_log /webapps/MyDjangoService/logs/nginx_error.log;

    location /static/ {
        alias   /webapps/MyDjangoService/static/;
    }

    location /media/ {
        alias   /webapps/MyDjangoService/media/;
    }

    location / {
        if (-f /webapps/MyDjangoService/maintenance_on.html) {
            return 503;
        }

        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto https;
        proxy_set_header Host $http_host;
        proxy_redirect off;

        # Try to serve static files from nginx, no point in making an
        # *application* server like Unicorn/Rainbows! serve static files.
        if (!-f $request_filename) {
            proxy_pass http://MyDjangoService_wsgi_server;
            break;
        }
    }

    # Error pages
    error_page 500 502 504 /500.html;
    location = /500.html {
        root /webapps/MyDjangoService/django/src/MyDjangoService/templates/;
    }

    error_page 503 /maintenance_on.html;
    location = /maintenance_on.html {
        root /webapps/MyDjangoService/;
    }
}

Answer 1

Your issue is in the following line:

        proxy_set_header X-Forwarded-Proto https;

Here you unconditionally set the X-Forwarded-Proto header to the value https. Your WSGI server will interpret this to mean that your site is running behind https. Django then does a strict referrer check, and sees that the protocol in the referrer domain is http instead of https. Because this can be a security issue, Django rejects the request.

You should either remove this line, or change it to use the correct value. You can use the $scheme variable for this:

    proxy_set_header X-Forwarded-Proto $scheme;

Answer 2

If for some reason you want or need to have https unconditionally forwarded and you are working in a localhost, get past the CSRF error on the admin page by putting

CSRF_TRUSTED_ORIGINS = ["127.0.0.1"]       
SECURE_PROXY_SSL_HEADER = ('HTTP_X_FORWARDED_PROTO', 'https')

in your settings.py file and using https://127.0.0.1:443/admin in your browser