Rookv1.18.8
Email is sent weekly only when there are changes to check; critical security advisories are sent immediately.
Important fixes6
Configured exporter options included in the
rook-ceph-clusterHelm chart templateAffectsrook-ceph-clusterHelm chart installations withmonitoring.exporteroptions configured(Inferred)When the chart generated a
CephClusterresource, configured node exporter options supported by the operator, such ashostNetwork, were omitted.- helm: Include exporter options in CephCluster (#16745, @michaeltchapman)
github.com/rook/rook/releases/tag/v1.18.8Rook CSI plugin resource assignments corrected
AffectsRook-managed RBD, CephFS, and NFS CSI drivers withControllerPluginandNodePluginresource settings.(Inferred)Previously, the settings were reversed: for example, the RBD controller received plugin resources while the node plugin received provisioner resources.
- csi: ControllerPlugin/NodePlugin resource settings were reversed (#16735, @swills)
github.com/rook/rook/releases/tag/v1.18.8Toolbox
ceph.confnow includesrook-config-overrideAffectsToolbox deployments with a non-emptyrook-config-overrideConfigMap(Inferred)When deployments relied on the ConfigMap for custom settings, the toolbox ignored it, leaving
ceph.confwith onlymon_hostand keyring settings and making diagnostic commands that need custom parameters difficult to run. The override is merged again when mon endpoints change.- toolbox: Merge rook-config-override ConfigMap into ceph.conf (#16731, @mheler)
github.com/rook/rook/releases/tag/v1.18.8PDB node-drain handling for snaptrim PG states fixed
AffectsCeph clusters with PGs insnaptrimorsnaptrim_waitstates(Inferred)When a node drain is evaluated, PG states matching
activefollowed by one or more ofclean,deep,scrubbing,snaptrim, orsnaptrim_waitare treated as healthy, so they do not block the drain.- osd: Allow snaptrimp and snaptrip_wait PGs by the PDBs during node drains (#16713, @sp98)
github.com/rook/rook/releases/tag/v1.18.8rook-ceph-cluster
HTTPRoutepath defaults fixedAffectsdashboard and CephObjectStoreHTTPRoutes withpathTypeunset(Inferred)Previously, the chart rendered the invalid
Prefixvalue whenpathTypewas unset. It now rendersPathPrefix.- helm: Fix default pathType for HTTPRoute in the rook-ceph-cluster chart (#16724, @fancl20)
github.com/rook/rook/releases/tag/v1.18.8Conflict retries for RADOS namespace and subvolume group status updates
AffectsCephBlockPoolRadosNamespaceandCephFilesystemSubVolumeGroupresources(Inferred)Previously, RADOS namespace status updates could fail and require a Rook operator pod restart, while subvolume group status updates had no conflict retries. Now, conflicting status updates are retried using the default policy, and an error is logged if retries still fail.
- namespace: Add retryOnConflict when updating status (#16661, @subhamkrai)
github.com/rook/rook/releases/tag/v1.18.8
Enhancements1
Ceph Tentacle added to supported versions
- core: Add support for ceph tentacle (#16501, @subhamkrai)
github.com/rook/rook/releases/tag/v1.18.8
Other changes2
Squid-to-Tentacle upgrade test added
- core: Add support for ceph tentacle (#16501, @subhamkrai)
github.com/rook/rook/releases/tag/v1.18.8RADOS namespace reconciliation waits for
CephBlockPoolreadiness- pool: Retry if pool status is empty in the rados namespace controller (#16705, @parth-gr)
github.com/rook/rook/releases/tag/v1.18.8
Summaries are produced by AI, so they can be wrong, and they are provided without warranty.