39

I am running a test django server on aws and I just installed django-userena and when I try to signup a user upon clicking submit, I get the following message:

relation "django_site" does not exist LINE 1: ..."django_site"."domain", "django_site"."name" FROM "django_si...

I am not really sure what went wrong here. I did some researching and added " 'django.contrib.sites'," to my installed apps, but I am still getting the error. I will there is an extra step I am missing. Any suggestions or advice?

RedShirt
  • 855
  • 2
  • 20
  • 33

13 Answers13

54

I recently ran into this issue (Django 1.8.7) even with SITE_ID = 1 in my settings. I had to manually migrate the sites app before any other migrations:

./manage.py migrate sites
./manage.py migrate
Franey
  • 4,164
  • 2
  • 16
  • 18
  • 2
    This did the trick! I already had `SITE_ID = 1` so the highest-voted answer didn't work for me – Hamman Samuel Dec 22 '15 at 15:24
  • 1
    I agree with you @HammanSamuel I have seen many solutions which never worked for me but I kept wondering why they were accepted. On a second thought I felt SITE_ID=1 used to work in the older versions of Djangon. However this did the Magic on the the lastest Django 1.9 as at the time of this comment – Paullo Jan 08 '16 at 06:05
  • Had this issue also, what fixed it for me was updating django_livesettings to the latest version. Were upgrading to django 1.8.12 when my tests stopped working and blamed `FOREIGN KEY (`site_id`) REFERENCES `django_site` (`id`)` when running the migrations for test database. – Gesias Apr 19 '16 at 15:13
  • I had a similar problem recently though it was when running tests to this solution didn't work. In the end I discovered that I was missing the 'allauth.socialaccount' app from my installed apps. I don't know why it suddenly started happening though. – Simon May 06 '21 at 06:19
11

You may be calling a site object before creating site model(before syncdb or migrate)

ex: site = Site.objects.get(id=settings.SITE_ID)

Geo Jacob
  • 5,909
  • 1
  • 36
  • 43
  • 2
    Good catch, thanks. On my dev site I enabled the sites framework, got some code working, everything fine. Pushed it to the live site and got this error. I had to comment out the Sites-related code, push to live, run migrations, uncomment the Sites-related code, and push to live again. Phew! – Phil Gyford Jul 08 '16 at 10:51
  • @PhilGyford A real pain! – joshlsullivan Dec 05 '18 at 16:12
  • That's quite a good point, leads simply to: **Always** read the stacktrace **carefully** :-D – normic Aug 21 '20 at 17:42
10

I have the same problem and fixed it like this:

  1. add SITE_ID=1 into settings.py

  2. run this command :

    python manage.py migrate
vabada
  • 1,738
  • 4
  • 29
  • 37
Saeed Ghareh Daghi
  • 1,164
  • 1
  • 13
  • 21
3

I got this error while working with django-cookiecutter, django-allauth and django-rest-auth

I literally spent 5 hours pulling my hair out. Eventually gave in and started to comment out bit by bit

What worked for me was commenting out both pre-configured url paths (they come with cookiecutter Django):

# User management
path("users/", include("yourapp.users.urls")),
path("accounts/", include("allauth.urls")),

After that migrations worked.

I uncommented it and my app has worked ever since. It was only for the initial migration

Hope it helps someone!

Raf Rasenberg
  • 534
  • 2
  • 14
  • 27
2

Going to leave this here for future me:

python manage.py makemigrations allauth

This worked for me, I forgot why, took me too long to figure out how I fixed this the first time

Edit: makemigrations sometimes doesnt make 3rd party things like allauth which some of my projects use, so I have to specify those ones

George Sferopoulos
  • 31
  • 3
2

if you are getting this error when deploying you django app to Heroku, make sure you have run:

heroku run python manage.py migrate

This worked for me

Fatima-Zahra Chriha
  • 21
  • 2
1

This issue might be caused by one of the apps you're using. If you check the traceback carefully, you might already find the delinquent.

I had those issues using django-debug-toolbarand zinnia.

If you are using the django-debug-toolbar this might be a solution:

Try following the steps for the explicit setup: http://django-debug-toolbar.readthedocs.org/en/1.2.2/installation.html#explicit-setup

Alternatively remove debug_toolbar from your INSTALLED APPS.

If that doesn't help or if another app is causing the issue, try to temporarily remove all imports (e.g. installed app, urls, custom views, settings), which are displayed in the traceback.

Tobias Lorenz
  • 1,426
  • 11
  • 17
1

A horrible code lead to this error for me. I had a global variable to get the current site

SITE = Site.objects.get(pk=1)

this was evaluated during migration and lead to the error.

Sandeep Balagopal
  • 1,943
  • 18
  • 28
1

I experienced the same problem with creating a new empty database for my project (which uses zinnia)

Running 'manage migrate site' before 'manage migrate' did not solve anything. It seems that the complete project was loaded before any table creating was done.

I resolved to catching the errors that importing the zinnia releated app produced.

e.g.: in the urls.py of the app

urlpatterns = None
app_name = 'something'

try:
    from .views import MyEntryCreate


    urlpatterns = [

    url(r'^blogentry/create/$',
        login_required(MyEntryCreate.as_view()),
        name='zinnia_entry-add'),

    ]
except Exception as e:
    logger.error(app_name+" Error urls: "+str(e))
    urlpatterns = []

Had to do something like that elsewhere in that app, and 'manage migrate' worked again.

Steckelfisch
  • 11
  • 2
1

I ran into the same problem today after pushing my code to heroku... So I figured out I had to migrate my sites before any other migration.

heroku run python manage.py migrate sites
heroku run python manage.py migrate
Oswald Tragger
  • 31
  • 2
0

I'm late, but I ran into the same issue with django v 1.11.

The issue was that I was rebuilding a model outside the normal def() and in a form() [I use the models for a choice] The traceback should have the .py file listed

e.g.

 File "filepath/views.py", line 67, in <module>
    some_variable = some_model.objects.get(name ='name')

So I had to comment it out to rebuild my migrations

user2723240
  • 803
  • 1
  • 13
  • 23
0

Posting this in case that someone lands on this problem and none of the answers above have worked.

This is an addition to one of the solutions that may or may not solve your problem:

The solution being: make sure that you do the migrate command first before anything else!

So in the usual case, you upload your code to the cloud, which could be AWS. You do the usual docker-compose commands right, well you should do exactly as the following first:

Do the following steps: docker-compose -f production.yml build then docker-compose -f production.yml run --rm django python manage.py migrate

And then after that you can now run the docker instance via docker-compose -f production.yml up or docker-compose -f production.yml up -d (search on google what that means)

--

If you've already run the instance before migrating, then just stop the docker instance that you're running and remove it. A handy tutorial for docker commands I've found is this: https://www.thegeekdiary.com/how-to-list-start-stop-delete-docker-containers/

Atheonth
  • 1
  • 2
-1

I just restarted my computer and problem disappeared :) (restarting docker-compose is not enough).

Arkadiusz Mazur
  • 53
  • 1
  • 2