Comment 1 for bug 1814916

Something is messed up because looking at the c-api logs, the MessagingTimeout happens before even 60 seconds have passed. The request starts here:

Jan 29 00:21:05.263845 ubuntu-bionic-ovh-bhs1-0002199912 <email address hidden>[12973]: INFO cinder.api.openstack.wsgi [req-d6667da6-d6a8-4912-8396-99d1a0b26631 req-10ef42dc-1cac-46c8-85b2-2166df98df94 tempest-VolumesExtendAttachedTest-619719536 tempest-VolumesExtendAttachedTest-619719536] PUT https://158.69.66.167/volume/v3/1a1deb76f71e4474aaa7a83958a7dd9d/attachments/5609ac83-2a2e-4207-b52b-4cea6371ecd2

and times out here:

Jan 29 00:21:05.738408 ubuntu-bionic-ovh-bhs1-0002199912 <email address hidden>[12973]: ERROR cinder.api.v3.attachments [req-d6667da6-d6a8-4912-8396-99d1a0b26631 req-10ef42dc-1cac-46c8-85b2-2166df98df94 tempest-VolumesExtendAttachedTest-619719536 tempest-VolumesExtendAttachedTest-619719536] Unable to update the attachment.: MessagingTimeout: Timed out waiting for a reply to message ID 51cf19ab9b244e1e9fe8598633187403

That's not even the total time that the RPC call from c-api to c-vol would take, so why are we getting a MessagingTimeout rather than the VolumeAttachmentNotFound exception in the API (which itself would be a 500 but would be a more clear error that we're probably racing with some kind of teardown issue in a test).