I'm getting this weird AttributeError on the app I'm currently working on. I'm developing using the development web server i.e. "runserver" command of django toolsets. Then I decided to test the application on apache+mod-wsgi and I persistently get this error though it works fine sometimes. So I think there must be something wrong with that piece of code,
so I decided to comment it out and see what happens. And YES, it still give me the same error (See 2nd picture). <-- NOT RELEVANT NOW. The AttributeError I'm getting on custom User model, even if it actually contains the classmethod get_by_type_and_id() is what I'm interested on.
Have you even seen like this one before? What do you think is causing this? I've followed the tutorial here to deploy it. Note though that User is not the built-in django User model. I think it's a "customized,stripped-down" implementation based from the django's Auth module.
Note that I have not gotten this error on my development using django's own development server. This only happens when I deployed the app on Apache+mod_wsgi.
More Info:
Django version == 1.2.5
Thanks! I'd really appreciate any kind of help.
First Picture:
2nd Picture:
Few things to notice:
Are you sure that you have get_by_type_and_id method for the User?
You set pasword instead of password (typo)
I think you have to indent line 60 in your second screenshot
After hours of diving into the code, I had a eureka moment. So I thought, maybe at some point django.contrib.auth.models.User model is being put into action by some django module used in the application. Because like I mentioned, this app's User model is customized and the application is not using django.contrib.auth.* at least directly. This was my suspicion because this kind of error is not specific to that particular attribute. Sometimes it gets through that point, but then another AttributeError would occur on some other view referring to other User model attributes. Then at one point I noticed that the attributes of User in the error messages was that of django.contrib.auth.models.User.
So I decided to "remove" django.contrib.auth.models on my python path, and there it showed, an error occurred in one of the modules I'm using specifically django.contrib.messages. I removed it, and I'm not seeing the AttributeError again. It turned out that django's Message model has a foreign_key to django.contrib.auth.models.User.
But then, I lost the flexibility of django message framework, especially when I'm using it on view with a HttpResponseRedirect i.e. no way to put useful messages in a template's context. or maybe there is? :)
UPDATE 2 (Pretty Sure Now):
Yes, I'm pretty sure now that django.contrib.messages is the one causing the error. I tried reproducing the error on another application which uses the customized User model i.e. not django.contrib.auth.models.User. I enabled django.contrib.messages module and produced basically the same error. I would like to know why is this happening on Apache+mod_wsgi, but not on django's own development server.
Related
Apologies if this question is not asked correctly but I actually can't pinpoint the problem. Hence the lack of code in the question itself. I'm struggling with this error so please go easy on me....
I was developing a meteor app locally without too much difficulty. It seems to run well on localhost. But then I decided to deploy it and have been running into some problems as outlined below. I have deployed the app "successfully" using both mup/DO and modulus.io (with compose hosting the DB in both instances) and whole thing sort of works, but...
You can register fine but when you try to log in the button doesn't work...press it again and you login but the usename and password are in the URL... how does this happen?
When you login you can begin by creating a supplier, then create products for that supplier. Only thing is the suppliers are saved to the DB and they are in the product count but they are nowhere to be seen in the middle section page.
Once you create a new supplier, the add new products for the existing suppliers is no longer accessible.
I am so confounded that I don't know what code to put up so I'm giving access to my codebase - it's on github here and this is the modulus site here. Go ahead and register and you can see for yourself.
You will easily be able to see the errors in the console when you start trying to log into the site so there is no point in posting them here.
Many, many thanks in advance guys.
I have reviewed your code and I don't think that the errors occur due to the deployment.
Here's a list with a few suggestions that should help you to fix your code:
In your /client/helpers/config.js file, you try to configure the behaviour of {{> loginButtons}}. That does not make any sense, since you do not have the accounts-ui package installed.
The /client/templates/includes/header.html file references with pathFor to homepage. This route is currently not available in your /lib/router.js.
Users are able to access the /loggedIn path even if they are not logged in. Furthermore, you always redirect users to this path if the submit form event in the register template occurs. This means, they can easily bypass the registration just by clicking on the submit button.
Watch your console logs. There are a lot of template helper exceptions.
Unfortunately, I could not check the login bug you described, because I received an exception when invoking the submit event. I recommend to use a rather defensive programming approach, you should at least check if the variable's value is not undefined and if it is, then you should handle those situations accordingly.
For example, in your /client/templates/includes/login.js file, you have the following code:
var userId = Meteor.userId();
var supplier = Suppliers.findOne({userId: userId});
var supplierId = supplier._id;
This will raise an exception if supplier is undefined.
All in all, you should rethink your release planning and deployment, since your app is far from working. Furthermore, please try to break your issue into chunks next time and provide a clear problem statement, because your question won't be useful to other readers without it.
I have a strange issue I've never seen before and I'm a bit stuck at where to go next.
We have a fairly standard Django app, with some forms using class based views, and we're finding that state is getting mixed between forms in strange ways.
For example, User A is on the site and fills out form X and sets their postcode.
User B goes to another form altogether, form Y, but the postcode field is already populated with User A's postcode.
The fields share the same name, but are on totally independent forms. The field in this example is a standard Django CharField.
I haven't managed to replicate this behaviour using the Django dev server locally from a little trying. I've upgraded gunicorn to latest version, I've set it to use a single worker, I've disabled caching everywhere in the Django project that I can find.
It's stuck on Django 1.3.7, and it's not a trivial app to upgrade to a supported version. It's running with gunicorn 18 behind nginx.
I've noted the Django mentions of thread safety in class based views, but I don't see us passing any mutable objects to the view, and it's all a fairly plan view and ModelForm. Django's session store is the default (it was previously something else but I removed that to use the Django default which I believe is the db).
Any ideas where to go next to try and figure this out?
Thanks!
Figured it out - turns out it was to do with the behaviour of get_initial in 1.3, which was changed in Django 1.4
https://docs.djangoproject.com/en/dev/releases/1.4/#formmixin-get-initial-returns-an-instance-specific-dictionary
We changed our get_initial to use a copy of the dict rather than the default implementation, and now all is well :)
I have a weird bug with django sessions in my app: some times (about 10 times for ~20000 per day) session information for user is erased. I traced it via log files: at page A there is information for user's session, after it he submits the form and at the next page his session is empty. I tried two types of storage: memcached+db and db only and this problem is for both of them. I tried to reproduce these scenarios, but all works as expected, as I said, it happens very rare. I also checked that this problem exists for different users, and for them is doesn't reproduce each time. I don't have any ideas how to catch the root cause and I don't know what else post here as a description. If someone has any ideas, please let me know. If it is important, I'm running my app with django 1.2 + FastCGI.
Thanks!
UPD: I checked and see that session key from uses is not changed during two sequential requests, at first request there is an actual session state, and at second session variables are relaced with empty.
As a way to debug this problem, I would subclass the standard Django session middleware (or whatever you're currently using):
django.contrib.sessions.middleware.SessionMiddleware
and wrap process_request and (probably more importantly) process_response in some extra logging. Then install your subclassed session middleware in the MIDDLEWARE_CLASSES, rather than the stock Django one.
You could also validate that session.save() has actually committed its changes by attempting to read it back. It could be that the problem lies in session-state serialisation, and it's failing on a particular key or value that you're attempting to store.
None of this will fix your problem, but it might help you to establish what's going on.
As #Steve Mayne mentioned, it would be good to do some logging on the sessions middleware and sessions model save method. That's something I'd start with.
In addition I'd like to say that this could be a database related issue, especially if you're using MySQL database backend for sessions. You can check the log for database locks and other concurrency issues. I had to deal with similar issues before and the solution is clear: optimization and additional performance.
If you have some specific application middleware, you can check for functionality that interferes with Django sessions. Such parallel operations can cause problems, if not implemented properly.
Another thing I would do is to upgrade to the latest stable release of Django and migrate to a mod_wsgi setup.
On the web page, I get the following error:
FieldError at /foo/bar/
Cannot resolve keyword 'foos' into field. Choices are: __unused__, [snip]
The problem code is
User.objects.filter(foos__name='bar')
When I run this in the shell, it works and I get a recordset:
>>> User.objects.filter(foos__name='bar')
[<User: JordanReiter>]
But on the webpage I get the exception above.
I've never run into this issue before and I wonder if I'm missing something?
Update
Based on doing a diff between the "Choices are: ..." on the web and in the shell it appears there are 7 fields that are available in the shell that aren't available if I do the query on the web. They appear to be ordinary ForeignKey fields pointing to User, with no difference from the other fields that work.
Tested So Far
INSTALLED_APPS are identical for both settings
runserver version also works (as would be expected)
User used is identical in both cases and is django.contrib.auth.models.User
related names for the shell's User and the web app's User are definitely different. User._meta.get_all_related_objects() in the shell displays around 7 more related fields than the if I dump that from the web app.
values for settings are also basically identical (one has the TEST_XYZ settings but those don't affect anything)
This is only sort-of an answer. Turns out the reason that the field is not available is because all of the installed apps' models do not correctly load in when the app is first compiled, so the app permanently believes certain fields don't exist, even after they recognize the models for those fields do exist. It appears to be related to this earlier, equally confusing problem:
SO: Internal Server error on the first request (and only the first request) after server reload
Oh, and so how I fixed it was to change areas of code that imported the models that the web server couldn't find on the first request. Somehow, omitting those meant that the server would recognize that the fields in question were present. Super weirdness!
Reverse foreign keys are only added to models if the app they are in is in INSTALLED_APPS. I know you say that your settings.py is the same as in development, but somehow it is the case that your foo app is in INSTALLED_APPS locally, but not on the server. Do you have a local_settings override perhaps?
I need to have a status message that is set while processing an initial request appear after a redirect occurs. A pretty normal thing to need to do but I'm unclear how to use the session object to do this in Django. I know there's a plugin someone made: https://github.com/danielfm/django-flash/wiki
Is that the recommended way of handling this situation in Django or can something simpler be done with the default install of Django?
Looks like this is part of django as of 1.2:
http://docs.djangoproject.com/en/dev/ref/contrib/messages/
There is no standardized (in Django) method yet for session messages.
Here's a good discussion of the issues: Towards a Standard for Django Session Messages
Besides the django-flash project you already found there is a least one other one: django-notify.