API returns UnknownError: An unknown error has occurred. Please try your request again. when describing instances

[Kevin Bringard]

When describing instances the following stack occurs:

2011-02-10 22:04:07,814 ERROR nova.api [OSE5QAE9DVBA9BDUDBPY user user] Unexpected error raised: Parent instance <FixedIp at 0x2ec4510> is not bound to a Session; lazy load operation of attribute 'network' cannot proceed
(nova.api): TRACE: Traceback (most recent call last):
(nova.api): TRACE: File "/usr/lib/python2.6/site-packages/nova-2011.1-py2.6.egg/nova/api/ec2/__init__.py", line 317, in __call__
(nova.api): TRACE: result = api_request.invoke(context)
(nova.api): TRACE: File "/usr/lib/python2.6/site-packages/nova-2011.1-py2.6.egg/nova/api/ec2/apirequest.py", line 128, in invoke(nova.api): TRACE: result = method(context, **args)(nova.api): TRACE: File "/usr/lib/python2.6/site-packages/nova-2011.1-py2.6.egg/nova/api/ec2/cloud.py", line 642, in describe_instances(nova.api): TRACE: return self._format_describe_instances(context, **kwargs)(nova.api): TRACE: File "/usr/lib/python2.6/site-packages/nova-2011.1-py2.6.egg/nova/api/ec2/cloud.py", line 649, in _format_describe_instances
(nova.api): TRACE: return {'reservationSet': self._format_instances(context, **kwargs)}
(nova.api): TRACE: File "/usr/lib/python2.6/site-packages/nova-2011.1-py2.6.egg/nova/api/ec2/cloud.py", line 690, in _format_instances
(nova.api): TRACE: floating_addr = fixed['floating_ips'][0]['address']
(nova.api): TRACE: File "/usr/lib/python2.6/site-packages/nova-2011.1-py2.6.egg/nova/db/sqlalchemy/models.py", line 74, in __getitem__
(nova.api): TRACE: return getattr(self, key)
(nova.api): TRACE: File "/usr/lib/python2.6/site-packages/SQLAlchemy-0.6.6-py2.6.egg/sqlalchemy/orm/attributes.py", line 168, in __get__
(nova.api): TRACE: instance_dict(instance))
(nova.api): TRACE: File "/usr/lib/python2.6/site-packages/SQLAlchemy-0.6.6-py2.6.egg/sqlalchemy/orm/attributes.py", line 388, in get
(nova.api): TRACE: value = callable_(passive=passive)
(nova.api): TRACE: File "/usr/lib/python2.6/site-packages/SQLAlchemy-0.6.6-py2.6.egg/sqlalchemy/orm/strategies.py", line 599, in __call__
(nova.api): TRACE: (mapperutil.state_str(state), self.key)
(nova.api): TRACE: DetachedInstanceError: Parent instance <FixedIp at 0x2ec4510> is not bound to a Session; lazy load operation of attribute 'network' cannot proceed
(nova.api): TRACE:
2011-02-10 22:04:07,816 ERROR nova.api [OSE5QAE9DVBA9BDUDBPY user user] Environment: {"HTTP_ACCEPT": "*/*", "CONTENT_TYPE": "application/x-www-form-urlencoded", "SCRIPT_NAME": "/services/Cloud", "REQUEST_METHOD": "POST", "HTTP_HOST": "10.1.170.32:8773", "PATH_INFO": "", "SERVER_PROTOCOL": "HTTP/1.0", "CONTENT_LENGTH": "-1", "HTTP_CONNECTION": "close", "SERVER_NAME": "10.1.170.32", "REMOTE_ADDR": "10.46.200.84", "wsgi.url_scheme": "http", "SERVER_PORT": "8773", "GATEWAY_INTERFACE": "CGI/1.1"}
2011-02-10 22:04:07,816 ERROR nova.api [OSE5QAE9DVBA9BDUDBPY user user] Environment: {"HTTP_ACCEPT": "*/*", "CONTENT_TYPE": "application/x-www-form-urlencoded", "SCRIPT_NAME": "/services/Cloud", "REQUEST_METHOD": "POST", "HTTP_HOST": "10.1.170.32:8773", "PATH_INFO": "", "SERVER_PROTOCOL": "HTTP/1.0", "CONTENT_LENGTH": "-1", "HTTP_CONNECTION": "close", "SERVER_NAME": "10.1.170.32", "REMOTE_ADDR": "10.46.200.84", "wsgi.url_scheme": "http", "SERVER_PORT": "8773", "GATEWAY_INTERFACE": "CGI/1.1"}
2011-02-10 22:04:07,817 ERROR nova.api [OSE5QAE9DVBA9BDUDBPY user user] UnknownError: An unknown error has occurred. Please try your request again.
2011-02-10 22:04:07,817 ERROR nova.api [OSE5QAE9DVBA9BDUDBPY user user] UnknownError: An unknown error has occurred. Please try your request again.
2011-02-10 22:04:07,817 INFO nova.api [OSE5QAE9DVBA9BDUDBPY user user] 0.220182s 10.46.200.84 POST None:None 400 [None] application/x-www-form-urlencoded text/xml
2011-02-10 22:04:07,817 INFO nova.api [OSE5QAE9DVBA9BDUDBPY user user] 0.220182s 10.46.200.84 POST None:None 400 [None] application/x-www-form-urlencoded text/xml

It only seems to happen with around 7 or more instances running per project, and happens much more frequently with users who do not have is_admin = 1.

Vishy provided the following patch, which appears to fix the issue:

http://pastie.org/1550608