Addressed issues¶
The following issues have been addressed in the MOSK 22.4 release:
[25349][Update] Fixed the issue causing MOSK cluster update failure after an OpenStack controller node replacement.
[26278][OpenStack] Fixed the issue with
l3-agent
being stuck in theNot ready
state and routers not being initialized properly during Neutron restart.[25447][OpenStack] Fixed the issue that caused a Masakari instance evacuation to fail if an encrypted volume was attached to a node.
[25448][OpenStack] Fixed the issue that caused some Masakari instances to get stuck in the
Rebuild
orError
state when being migrated to a new OpenStack compute node during host evacuation. The issue occurred on OpenStack compute nodes with a large number of instances.[22930][OpenStack] Fixed the issue wherein Octavia load balancers provisioning, and, occasionally, the listeners or pools associated with these load balancers got stuck in the
ERROR
,PENDING_UPDATE
,PENDING_CREATE
, orPENDING_DELETE
state.[25450][OpenStack] Implemented the capability to enable
trusted
mode for SR-IOV ports.[25316][StackLight] Introduced projects filtering by a domain name for the
default
domain to fix the issue wherein a wrong project was chosen by name in case of multiple projects with the same names.[24376][Ceph] Implemented the capability to parametrize the RADOS Block Device (RBD) device map to avoid Ceph volumes being unresponsive due to a disabled cyclic redundancy check (CRC) mode. Now you can use the
rbdDeviceMapOptions
field in the Ceph pool parameters of theKaaSCephCluster
CR to specify custom RBD map options to use withStorageClass
of a corresponding Ceph pool. For details, see Pool parameters.[28783] [Ceph] Fixed the issue causing Ceph conditon stuck in absence of the Ceph cluster secrets information. If you applied the workaround to your MOSK 22.3 cluster before the update, remove the
version
parameter definition fromKaaSCephCluster
after the managed cluster update because the Ceph cluster version in MOSK 22.4 updates to 15.2.17.