I am still working on analysis of this bug, but it seems to be easily reproducable and does not seem to be (directly) related to persistent storage being wiped. I currently have the following scenario:
1. Boot the network node (failover node)
2. Notice the following in /var/log/neutron/l3-agent.log
4. Notice the following lines in /var/log/neutron/l3-agent.log
WARNING neutron.agent.l3.agent [-] Info for router a94b7f00-36ba-4944-bc48-025473758224 was not found. Performing router cleanup
INFO neutron.agent.l3.ha [-] Router a94b7f00-36ba-4944-bc48-025473758224 transitioned to backup
6. The system now performs as expected
5. Now, stop the l3-agent
I am still working on analysis of this bug, but it seems to be easily reproducable and does not seem to be (directly) related to persistent storage being wiped. I currently have the following scenario:
1. Boot the network node (failover node) neutron/ l3-agent. log
2. Notice the following in /var/log/
ERROR oslo_service. periodic_ task [req-8b6c1048- 1a4a-484c- bdf6-e069f478a1 44 - - - - -] Error during L3NATAgentWithS tateReport. periodic_ sync_routers_ task
3. Fix the problem by removing the router from the l3-agent
$ neutron l3-agent- router- remove <agent id> <router id>
4. Notice the following lines in /var/log/ neutron/ l3-agent. log
WARNING neutron. agent.l3. agent [-] Info for router a94b7f00- 36ba-4944- bc48-0254737582 24 was not found. Performing router cleanup 36ba-4944- bc48-0254737582 24 transitioned to backup
INFO neutron.agent.l3.ha [-] Router a94b7f00-
6. The system now performs as expected
5. Now, stop the l3-agent
# systemctl stop neutron-l3-agent
6. Start the l3-agent
# systemctl start neutron-l3-agent
7. Notice the same error as (2) re-appears