fix test_allowed_address_pair test case to fix race condition during network deletion
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
tempest |
Fix Released
|
Undecided
|
Gayathri Devi Kathiri |
Bug Description
"test_allowed_
Details: {'type': 'SubnetInUse', 'message': 'Unable to complete operation on subnet e6615d0f-
Details: {'type': 'NetworkInUse', 'message': 'Unable to complete operation on network 863441e6-
tempest.
tempest.
tempest.
tempest.
Looks like subnet/network delete is triggered while port being deleted.
This rare scenario is observed when whole tempest repo test suite is run.
Since the tempest test case failed at cleanup with error indicating that ports are still in use on the network, logged into the environment where test case failed and found that port delete triggered by the tempest is successful. So, there are no ports associated with the network/subnet after test case failure.
Neutron deleted the port as triggered by the tempest. Network/subnet existed in the environment since it failed at the cleanup of networks/subnet.
Tempest code calls the delete_port method but doesn't wait whether port is deleted which could lead to a rare situation when port is being deleted but tempest is already trying to delete the network.
Changed in tempest: | |
assignee: | nobody → Gayathri Devi Kathiri (gayathridevikathiri) |
Changed in tempest: | |
status: | New → In Progress |
description: | updated |
There has been the same problem with other network tests recently: https:/ /bugs.launchpad .net/tempest/ +bug/1893487