Bert, thanks for the offer to help.
So Nova uses this FaultWrapper middleware (https://github.com/openstack/nova/blob/3b2895b960630934f827918d50577f9167110cab/nova/api/openstack/__init__.py#L106) to intercept all exceptions and problems so it can re-wrap them in its Fault class. I wonder if something changed that stopped calling the Fault instance. It seems like we're expecting all of the work that is causing this to happen in the Fault's __call__ (https://github.com/openstack/nova/blob/2bc8128d7793cc72ca2e146de3a092e1fef5033b/nova/api/openstack/wsgi.py#L1061) method and I wonder if something stopped calling it.
Note also that the FaultWrapper middleware *returns* the exception instead of re-raising the exception wrapped in the Fault. Could either of those (or both) be related to this problem?
Bert, thanks for the offer to help.
So Nova uses this FaultWrapper middleware (https:/ /github. com/openstack/ nova/blob/ 3b2895b96063093 4f827918d50577f 9167110cab/ nova/api/ openstack/ __init_ _.py#L106) to intercept all exceptions and problems so it can re-wrap them in its Fault class. I wonder if something changed that stopped calling the Fault instance. It seems like we're expecting all of the work that is causing this to happen in the Fault's __call__ (https:/ /github. com/openstack/ nova/blob/ 2bc8128d7793cc7 2ca2e146de3a092 e1fef5033b/ nova/api/ openstack/ wsgi.py# L1061) method and I wonder if something stopped calling it.
Note also that the FaultWrapper middleware *returns* the exception instead of re-raising the exception wrapped in the Fault. Could either of those (or both) be related to this problem?