Considering the following configuration:
$ovn-nbctl acl-list sw01 from-lport 100 (inport == "sw01-port1" && udp.dst == 5201) allow-related [after-lb] from-lport 10 (inport == "sw01-port1" && udp) allow-related [after-lb] $ovn-nbctl list acl _uuid : e440336a-84d3-4a6d-95a9-edd1db1c3631 action : allow-related direction : from-lport external_ids : {} label : 0 log : false match : "inport == \"sw01-port1\" && udp" meter : [] name : [] options : {apply-after-lb="true"} priority : 10 sample_est : ac6a6efc-a2e0-4d68-b5f8-8cd91113e554 sample_new : 5cdad2ab-4390-4772-ac40-74aa2980c06e severity : [] tier : 0 _uuid : 85ef08d7-aacc-41d7-b808-6ab011edd753 action : allow-related direction : from-lport external_ids : {} label : 0 log : false match : "inport == \"sw01-port1\" && udp.dst == 5201" meter : [] name : [] options : {apply-after-lb="true"} priority : 100 sample_est : 143ce7e2-fd13-4d5e-930c-133d5cf87d0d sample_new : 1d1a0a05-2a8a-4c72-ad35-77d7e2908183 severity : [] tier : 0
If the priority-100 ACL is removed, a long standing udp connection with destination port 5201 will hit the 10-priority ACL however ovn-controller will continue sampling the existing connection with the observationPointID associated to the removed ACL. In order to fix the issue we should recommit the observationPointID to the connection tracking table.
- links to
-
RHBA-2025:144760 ovn24.09 bug fix and enhancement update