User has no config

[someconservator]

NetBox version

v.3.1.8

Python version

3.9

Steps to Reproduce

1. Click "log in"
2. Enter credentials
3. Click "sign in"

Expected Behavior

Correct login

Observed Behavior

<class 'django.contrib.auth.models.User.config.RelatedObjectDoesNotExist'>

User has no config.

Python version: 3.9.7
NetBox version: 3.1.8

[jeremystretch]

Thank you for opening a bug report. Unfortunately, the information you have provided is not sufficient for someone else to attempt to reproduce the reported behavior. Remember, each bug report must include detailed steps that someone else can follow on a clean, empty NetBox installation to reproduce the exact problem you're experiencing. These instructions should include the creation of any involved objects, any configuration changes, and complete accounting of the actions being taken. Also be sure that your report does not reference data on the public NetBox demo, as that is subject to change at any time by an outside party and cannot be relied upon for bug reports.

[someconservator]

Excuse me, but what other steps need to be specified? This is exactly what happens to me: I try to log in and I get a window with this error.

[jeremystretch]

Per the bug report template:

Describe in detail the exact steps that someone else can take to reproduce this bug using the current stable release of NetBox.

You have not provided any such detail. Obviously merely logging into NetBox does not reproduce the problem, as you can verify for yourself here. If you'd like someone to look into this, you'll need to spend some time determining how to reproduce it.

[someconservator]

During today I've narrowed down the possible issues: a fresh installation of netbox (by the way, this is centos8.5, and nginx1.20, and with ldap) does indeed work fine. My netbox installation is an upgraded installation from version 3.1. As soon as i dump its database to a fresh installation i immediately get the same error. I have a suspicion that the update script did not make some necessary migrations (while the data itself is present in the database, I checked)

[someconservator]

During today I've narrowed down the possible issues: a fresh installation of netbox (by the way, this is centos8.5, and nginx1.20, and with ldap) does indeed work fine. My netbox installation is an upgraded installation from version 3.1. As soon as i dump its database to a fresh installation i immediately get the same error. I have a suspicion that the update script did not make some necessary migrations (while the data itself is present in the database, I checked)

[someconservator]

When i try to login in DEBUG mode in test server, i get on console:
[22/Feb/2022 08:22:22] "POST /login/ HTTP/1.1" 302 0
Internal Server Error: /
Traceback (most recent call last):
File "/opt/netbox-3.1.8/venv/lib64/python3.9/site-packages/django/core/handlers/exception.py", line 47, in inner
response = get_response(request)
File "/opt/netbox-3.1.8/venv/lib64/python3.9/site-packages/django/core/handlers/base.py", line 181, in _get_response
response = wrapped_callback(request, *callback_args, **callback_kwargs)
File "/opt/netbox-3.1.8/venv/lib64/python3.9/site-packages/django/views/generic/base.py", line 70, in view
return self.dispatch(request, *args, **kwargs)
File "/opt/netbox-3.1.8/venv/lib64/python3.9/site-packages/django/views/generic/base.py", line 98, in dispatch
return handler(request, *args, **kwargs)
File "/opt/netbox-3.1.8/netbox/netbox/views/init.py", line 136, in get
changelog_table = ObjectChangeTable(changelog, user=request.user)
File "/opt/netbox-3.1.8/netbox/utilities/tables.py", line 68, in init
selected_columns = user.config.get(f"tables.{self.class.name}.columns")
File "/opt/netbox-3.1.8/venv/lib64/python3.9/site-packages/django/utils/functional.py", line 247, in inner
return func(self._wrapped, *args)
File "/opt/netbox-3.1.8/venv/lib64/python3.9/site-packages/django/db/models/fields/related_descriptors.py", line 421, in get
raise self.RelatedObjectDoesNotExist(
django.contrib.auth.models.User.config.RelatedObjectDoesNotExist: User has no config.

How I can further understand, what gives me such an error?

[someconservator]

I found!
When migrating from an older version, the users_userconfig table remains empty:
netboxdb=> select * from users_userconfig;
id | data | user_id
----+------+---------
(0 rows)

when i manually do:
netboxdb=> INSERT INTO users_userconfig values (1,'{}',1);

this error disappears.

hence the questions:
why does this happen?
how best to fix this?
will this continue to happen when a new user is added?

[Jacq]

In my case it was a fresh install of Netbox version 0b508e83bd3 (v3.2.2), exactly the same issue, we could complete the first login of the admin user and we had to create manually the users_userconfig record as @someconservator.

[KyleKaniecki]

This happens with me as well

<class 'django.contrib.auth.models.User.config.RelatedObjectDoesNotExist'>

User has no config.

Python version: 3.9.5
NetBox version: 3.2.3

Logging in with admin/admin on a fresh install

[Guruleenyc]

Same issue as well.

[prnzngr]

I have the same issue. Any news?

[Mikesco3]

same issue here...

[cpmills1975]

I've just hit the same issue. In my case, I'm trying to build a NetBox installation running on Kubernetes on a cluster of Raspberry Pi's in my home lab. So, under the hood I am running the following containers:

• docker.io/tgenannt/netbox:v3.2.6 - the in development ARM64 version of netbox-docker
• ghcr.io/zcube/bitnami-compat/postgresql:14.4.0-debian-11-r10 - the ARM64 build of bitnami's Postgres 14
• ghcr.io/zcube/bitnami-compat/redis:7.0.3-debian-11-r14 - the ARM64 build of bitnami's Redis 7

It took some effort to get this environment up and running, partly because I think the instructions for the Kubernetes install are still based on NetBox 2.11 but, having got all the containers running, I'm now hitting the same issue/bug when I try and log in with my super user account admin/admin.

What isn't yet clear to me, and I'll try and work it out, is whether this is something to do with NetBox or something to do with the process that creates the super user in netbox-docker which I believe is somewhat different to the documented netbox install process. If there is a bug to fix, I guess I (we) need first to work out whose issues that bug should sit under!

[cpmills1975]

This is the code they are using (in docker-entrypoint.sh):

  ./manage.py shell --interface python <<END
from django.contrib.auth.models import User
from users.models import Token
if not User.objects.filter(username='${SUPERUSER_NAME}'):
    u=User.objects.create_superuser('${SUPERUSER_NAME}', '${SUPERUSER_EMAIL}', '${SUPERUSER_PASSWORD}')
    Token.objects.create(user=u, key='${SUPERUSER_API_TOKEN}')
END

[jeremystretch]

Is the user actually being created though, or it does it already exist? create_superuser() should work as expected (it's just a wrapper around save() on the instance).

>>> user4 = User.objects.create_superuser('user4')
>>> user4.config
<UserConfig: UserConfig object (9)>

[cpmills1975]

The user must have been created as this was a brand new install with a blank DB. The one thing that does come to mind is that the Kubernetes deployment brought up three replicas simultaneously - is there any way that may have caused some sort of race condition?

The log output shows that the relevant block of code executed, but perhaps it didn't - the super users' token didn't get created either.

Hmm... I'm now wondering whether the DB wasn't empty completely. I mentioned above that getting NetBox to run was a challenge due to the instructions being based on an old version. I hit some issues with redis authentication on the cache database. Is there any chance that could have perhaps interrupted the user creation process?

[jeremystretch]

Hard to speculate on what it could be. Even if it was a race condition, it should still be getting created in the database via some other instance.

The simplest solution might be to just add a check in NetBox's login function to ensure that a UserConfig exists for the user after authentication, and if not to automatically create one.

[jeremystretch]

I've opened #9741 to add a secondary check/creation for the UserConfig instance on login. Hopefully this should avoid the issue altogether.

[jeremystretch]

Just to reiterate from the thread above, I've opened #9741 to add an additional check for a missing UserConfig during login. Although we haven't been able to pinpoint the exact issue within Docker, this addition should avoid the issue entirely.

Thanks @cpmills1975 for all your work in tracking this down!

[Mikesco3]

I'm my case, I was trying to specify a user and password other than admin/admin in the docker compose yaml.
But then I destroyed the container and just used admin/admin to get it going and it worked.
So once in Netbox I created my super user and deleted admin.
I speculate it is something related to the script that creates the user. Is there a part that is hard coding the creation of an admin/admin or expecting a user called admin regardless of what the docker compose configuration says?

[cpmills1975]

Thanks @jeremystretch for looking in to this so quickly. I suspect my issue was an edge case, but your belt and braces approach may help others - it was hitting this error that lead me to this discussion.

So to dig further in to the root cause of this, I deleted the NetBox deployment (leaving the database untouched), changed the superuser account name, in my case in the Kubernetes configmap, but I guess the Docker compose file would be equivalent, then recreated the deployment and I can confirm that my second admin account was properly created along with a UserConfig.

[ed-ud]

This bug just started happening to me after upgrading our NetBox instance from v3.5.2 to v3.6.4.

The problem appears to be, when I create a new user, it doesn't add a new row to the users_userconfig table for that id. I tried creating and deleting several users before I finally figured this out. Manually adding an "empty" config row to users_userconfig with the new id appears to work around the problem.

[ed-ud]

@jeremystretch - should I open a new issue or did you want to revive the existing issue: #9741 ?

[kudlaty79bc]

Hello,
Have the same issue in 3.6.4 (new installation with database transfered from 3.6.3).
Using below command directly in database works
nbdb=> INSERT INTO users_userconfig values (1,'{}',1);
where "1" is ID and USER_ID

But that is definitely a bug.

[ed-ud]

I guess we'll need to create a new issue...

[Guruleenyc]

How do I fix this on a brand new install after trying to login ?