Description of problem: OVN controller network agents are not alive after turning off/turning on all controllers, see below (overcloud) [stack@undercloud-0 ~]$ openstack network agent list +--------------------------------------+------------------------------+--------------------------+-------------------+-------+-------+-------------------------------+ | ID | Agent Type | Host | Availability Zone | Alive | State | Binary | +--------------------------------------+------------------------------+--------------------------+-------------------+-------+-------+-------------------------------+ | cd0be7d2-5965-4bf2-aaaf-d69290558ff8 | OVN Controller agent | compute-1.localdomain | n/a | XXX | UP | ovn-controller | | 3f33c7ef-7226-430f-8e36-11851c2b3e6e | OVN Metadata agent | compute-1.localdomain | n/a | :-) | UP | networking-ovn-metadata-agent | | 862f5204-0b50-4412-8263-a08a7c762832 | OVN Controller Gateway agent | controller-0.localdomain | n/a | XXX | UP | ovn-controller | | 3f8bd2c6-49bc-4a7d-93fe-c31ef0be041e | OVN Controller Gateway agent | controller-1.localdomain | n/a | XXX | UP | ovn-controller | | 6636500c-c048-4e00-9dd4-9cd0e0440952 | OVN Controller agent | compute-0.localdomain | n/a | XXX | UP | ovn-controller | | 3fd9452b-bd0b-4dcf-9b44-712ddb98dbe8 | OVN Metadata agent | compute-0.localdomain | n/a | :-) | UP | networking-ovn-metadata-agent | | 84f405f4-03c8-4a0a-acb5-2beeead4f1c5 | OVN Controller Gateway agent | controller-2.localdomain | n/a | XXX | UP | ovn-controller | +--------------------------------------+------------------------------+--------------------------+-------------------+-------+-------+-------------------------------+ Version-Release number of selected component (if applicable): 15.0-RHEL-8/RHOS_TRUNK-15.0-RHEL-8-20190527.n.0 How reproducible: 100% Steps to Reproduce: Using an environment with 3 controllers. 1. Turn off all 3 controllers (I ran 'pcs cluster stop' then 'shutdown -h now' on each) 2. Turn on all 3 controllers and wait until they are up and ovn-dbs-bundle has a master node. Make sure 'sudo pcs status' shows all resources running properly. 3. Check 'openstack network agent list' Actual results: openstack network agent list shows that ovn controller agents are not alive it is not possible to run a new instance, it is created in ERROR state Expected results: ovn controller agents are alive It is possible to run new instances Additional info: The following error shown in ovn controller log 2019-06-18T13:24:13Z|00025|physical|ERR|No tunnel endpoint found for HA chassis in HA chassis group of port cr-lrp-e18b45c0-aa77-4908-a7e3-07dba18dc73a 2019-06-18T13:24:13Z|00026|binding|INFO|Releasing lport cr-lrp-7e9e8da9-687d-459e-8ba4-c0b29098254c from this chassis. Note: all instances that were created previously remain running and accessible. After restarting all ovn-controller containers network agents are alive and it is possible to run instances.
*** This bug has been marked as a duplicate of bug 1720947 ***