Addressed issues

The following issues have been addressed in the MOSK 22.3 release:

  • [23771][Update] Fixed the issue that caused connectivity loss due to a wrong update order of Neutron services.

  • [23131][Update] Fixed the issue that caused live migration to fail during an update of a cluster with encrypted storage. Now, you can perform live migrations using the following command:

    openstack --os-cloud osctl server migrate --live-migration <INSTANCE_ID>

  • [21790][Update] Fixed the issue wherein the Ceph cluster failed to update on a managed cluster with the Daemonset csi-rbdplugin is not found error message.

  • [23985][OpenStack] Fixed the issue that caused the federated authorization failure on the Keycloak URL update.

  • [23484][OpenStack] To prevent from ovsdb session disconnection causing port processing operations being blocked, configured the default timeouts for ovsdb.

  • [23297][OpenStack] Fixed the issue that caused VMs to be unaccessible through floating IP due to missing IPtables rules for floating IPs on the OpenStack compute DVR router.

  • [23043][OpenStack] Updated Neutron Open vSwitch to version 2.13 to fix the issue that caused broken communication between VMs in the same network.

  • [19065][OpenStack] Fixed the issue that caused Octavia load balancers to lose Amphora VMs after failover.

  • [23338][Tungsten Fabric] Fixed the issue wherein the Tungsten Fabric (TF) tools from the contrail-tools container did not work on DPDK nodes.

  • [22273][Tungsten Fabric] To avoid issues with CassandraCacheHitRateTooLow StackLight alerts raising for the tf-cassandra-analytics Pods, implemented the capability to configure file_cache_size_in_mb for the tf-cassandra-analytics or tf-cassandra-config Cassandra deployment. By default, this parameter is set to 512. For details, see Cassandra configuration.