Description of problem: After an OSP16.1 update (the update was performed from RHOS-16.1-RHEL-8-20200903.n.0 to RHOS-16.1-RHEL-8-20201007.n.0), controller-1 acts as master haproxy and controller-0 hosts master OVN DBs. Nodes have not been rebooted (this is different from the scenario from BZ1885592). Creation of network resources does not fail (this is different from BZ1885592 too). Network creation and deletion was performed multiple times successfully. However, openstack network agent list command fails 1/3 times. This is the failure message shown: (overcloud) [stack@undercloud-0 ~]$ openstack network agent list +--------------------------------------+------------------------------+---------------------------+-------------------+-------+-------+----------------+ | ID | Agent Type | Host | Availability Zone | Alive | State | Binary | +--------------------------------------+------------------------------+---------------------------+-------------------+-------+-------+----------------+ | 0866d9c3-2e61-4a49-8585-cd25ca7e59bf | OVN Controller Gateway agent | controller-1.redhat.local | | XXX | UP | ovn-controller | | 310c4d8e-7e70-42dd-b7e6-09c40f9b5c1c | OVN Controller Gateway agent | controller-2.redhat.local | | :-) | UP | ovn-controller | | 7aab32fb-5001-49be-9a50-5c5e1856855d | OVN Controller agent | compute-1.redhat.local | | :-) | UP | ovn-controller | | 1280a5b6-c18e-4cf9-a415-003effafca1c | OVN Controller Gateway agent | controller-0.redhat.local | | XXX | UP | ovn-controller | | ccf0ef5f-8258-4f1f-82ab-b1497522a816 | OVN Controller agent | compute-0.redhat.local | | :-) | UP | ovn-controller | +--------------------------------------+------------------------------+---------------------------+-------------------+-------+-------+----------------+ Logs show errors happen when controller-0's neutron-api handles the agent list request. The following errors are shown only on controller-0: 2020-10-08 16:41:15.172 34 WARNING neutron.db.agents_db [req-05155849-657f-4cd5-b9f5-2db4061cc246 - - - - -] Agent healthcheck: found 2 dead agents out of 5: Type Last heartbeat host OVN Controller Gateway agent 2020-10-08 16:41:15.167866 controller-1.redhat.local OVN Controller Gateway agent 2020-10-08 16:41:15.170482 controller-0.redhat.local pcs status shows no issues with OVN DBs containers status is up According to neutron-api logs, all processes were successfully connected to Neutron DBs after the OSP update. Version-Release number of selected component (if applicable): the update was performed from RHOS-16.1-RHEL-8-20200903.n.0 to RHOS-16.1-RHEL-8-20201007.n.0 How reproducible: only tried once and it failed Steps to Reproduce: 1. perform and osp16.1 update. After the overcloud update is completed, run 'openstack network agent list' several times (it failed 1/3 times)
*** This bug has been marked as a duplicate of bug 1788336 ***