I made locally some tests. This issue can be reproduced even locally but it happens once per about 20 or 30 times. I did it with command:
i=1; while true; do echo "========= Test $i starting =========="; tox -e dsvm-fullstack neutron.tests.fullstack.test_qos.TestBwLimitQoSOvs.test_bw_limit_qos_port_removed; RES=$?; i=$(( i + 1 )); if [ $RES -ne 0 ]; then break; fi; done
Problem is that I have no idea why it happens like that. From logs (openvswitch agent and test log, see attachments) it looks like specified "qos" and "queue" records are not deleted from ovs db. But DbDestroy Commands for both records are called properly, there is no errors in those transactions. Also strange thing is that after test is finished, those records are deleted from ovs.
I made locally some tests. This issue can be reproduced even locally but it happens once per about 20 or 30 times. I did it with command:
i=1; while true; do echo "========= Test $i starting =========="; tox -e dsvm-fullstack neutron. tests.fullstack .test_qos. TestBwLimitQoSO vs.test_ bw_limit_ qos_port_ removed; RES=$?; i=$(( i + 1 )); if [ $RES -ne 0 ]; then break; fi; done
Problem is that I have no idea why it happens like that. From logs (openvswitch agent and test log, see attachments) it looks like specified "qos" and "queue" records are not deleted from ovs db. But DbDestroy Commands for both records are called properly, there is no errors in those transactions. Also strange thing is that after test is finished, those records are deleted from ovs.