Exception on first syncdb with high replication / SQlite dev server

[wrwrwr]

WARNING:root:The rdbms API is not available because the MySQLdb library could not be loaded.
Creating tables ...
Traceback (most recent call last):
  File "manage.py", line 11, in <module>
    execute_manager(settings)
  File "/a/django/testapp/django/core/management/__init__.py", line 438, in execute_manager
    utility.execute()
  File "/a/django/testapp/django/core/management/__init__.py", line 379, in execute
    self.fetch_command(subcommand).run_from_argv(self.argv)
  File "/a/django/testapp/django/core/management/base.py", line 191, in run_from_argv
    self.execute(*args, **options.__dict__)
  File "/a/django/testapp/django/core/management/base.py", line 220, in execute
    output = self.handle(*args, **options)
  File "/a/django/testapp/django/core/management/base.py", line 351, in handle
    return self.handle_noargs(**options)
  File "/a/django/testapp/django/core/management/commands/syncdb.py", line 109, in handle_noargs
    emit_post_sync_signal(created_models, verbosity, interactive, db)
  File "/a/django/testapp/django/core/management/sql.py", line 190, in emit_post_sync_signal
    interactive=interactive, db=db)
  File "/a/django/testapp/django/dispatch/dispatcher.py", line 172, in send
    response = receiver(signal=self, sender=sender, **named)
  File "/a/django/testapp/django/contrib/contenttypes/management.py", line 20, in update_contenttypes
    content_types.remove(ct)
ValueError: list.remove(x): x not in list

Only happens sometimes (nearly always with one particular project), fairly hard to reproduce, feels like a concurrency issue (but disabling threading doesn't help) or something similar to this: https://code.djangoproject.com/ticket/7052. With testapp I get it most often after adding:

DATABASES['default']['DEV_APPSERVER_OPTIONS']
    = { 'high_replication' : True, 'use_sqlite': True, }

and using a specific set and order of INSTALLED_APPS:

'django.contrib.admin',
'django.contrib.auth',
'django.contrib.contenttypes',
'django.contrib.sessions',
'django.contrib.sites',
'djangotoolbox',
'autoload',
'dbindexer',
'djangoappengine',

Note that the order is actually significant -- swapping auth with contenttypes seems to fix the issue... Also note that for testapp frequency of this happening may be low (varies a lot depending on what exactly you have in settings), so to see it you'll likely need something like this:

for i in {1..100}; do rm -r .gaedata; python2.7 manage.py syncdb --noinput; done

[aburgel]

whats the value of ct that is not in the list?

does turning off high_replication help? maybe using high replication on the dev server does some eventual consistency emulation.

[wrwrwr]

Removing 'high_replication', 'use_sqlite' or any of the contrib apps does help.

When the exception occurs the 'content_types' list is empty while Django is trying to remove 'site'.

[aburgel]

i ran into this issue myself recently. i'm pretty sure its an eventual consistency issue.

i was able to work around it by temporarily changing the consistency policy, but its a bit of a hack. it also requires modifying the command code.

a better solution might be to always run commands with a strong consistency policy and then change to eventual consistency explicitly. like when running tests or the actual app.