ConfigMap pelagia-lcmconfig#
Pelagia controllers use a common configuration place that is the
pelagia-lcmconfig Configmap
object located in the same namespace as the Pelagia controllers. This Configmap is created by the Pelagia Helm chart.
Some parameters are applied from Helm chart values when you upgrade the chart,
others are not updated by the chart and must be edited directly in pelagia-lcmconfig.
Configuration options for Pelagia controllers#
If a parameter is not specified in pelagia-lcmconfig, it uses the code-based default value.
Options configurable through chart values#
The following Pelagia controller options are set in Pelagia Helm chart values during chart update.
The Default column lists the code-based default. The Helm chart may override it with its own default value. For the chart defaults, see Helm chart configuration.
Some parameters are written to pelagia-lcmconfig only under certain conditions, as noted in the Description column.
All DEPLOYMENT_* parameters are written only if cephDeployment.enabled is true.
If the chart omits a parameter, the code default applies.
| Parameter | Description | Default | Chart value |
|---|---|---|---|
| ROOK_NAMESPACE | Rook namespace. | "rook-ceph" |
lcmConfig.rookNamespace |
| DISK_DAEMON_API_PORT | Port for the disk daemon API. | "9999" |
lcmConfig.diskDaemonPortParameter |
| DISK_DAEMON_PLACEMENT_NODES_SELECTOR | Label for disk daemon placement. | "pelagia-disk-daemon=true" |
lcmConfig.diskDaemonNodeSelector |
| GATEWAY_API_ENABLED | Enable usage of the Gateway API. | "true" |
lcmConfig.gatewayAPIEnabled |
| GATEWAY_BASE_NAME | Name of the Gateway object used by default. Set only if lcmConfig.gatewayAPIEnabled is true and lcmConfig.gatewayName is not empty. |
"app-gateway" |
lcmConfig.gatewayName |
| GATEWAY_BASE_NAMESPACE | Namespace of the Gateway object used by default. Set only if lcmConfig.gatewayAPIEnabled is true and lcmConfig.gatewayNamespace is not empty. |
"openstack" |
lcmConfig.gatewayNamespace |
| RGW_PUBLIC_ACCESS_SERVICE_SELECTOR | Label of the service or proxy exposing RGW to public access. | "external_access=rgw" |
lcmConfig.rgwPublicAccessServiceSelector |
| KEEP_INGRESS | Deprecated. Enable support for Ingress usage. Will be removed in the following release due to NGINX Ingress retirement. Set only if lcmConfig.useIngress is true. |
"false" |
lcmConfig.useIngress |
| DEPLOYMENT_CEPH_RELEASE | Pin the Ceph release for the current setup. If empty, uses the latest available release for the current version. | "" |
cephRelease |
| DEPLOYMENT_NETPOL_ENABLED | Enable creation of network policy. | "true" |
cephDeployment.netpolEnabled |
| DEPLOYMENT_OPENSTACK_CEPH_SHARED_NAMESPACE | Namespace for the Openstack-Ceph communication and secrets sharing. Set only if cephDeployment.installSharedNamespace is true. |
"openstack-ceph-shared" |
cephDeployment.openstackSharedNamespace |
| DEPLOYMENT_LABEL_TO_EXCLUDE_CEPH_DAEMONSETS | Label for nodes where no Ceph daemons must be scheduled. Set only if the chart value is not empty. | "" |
lcmConfig.cephDaemonsetLabelExclude |
| DEPLOYMENT_DRAIN_REQUEST_LABEL_KEY | Label key marking a node as drained. Set only if the chart value is not empty. | "kaas.mirantis.com/lcm-drained" |
cephDeployment.drainRequestLabelKey |
| DEPLOYMENT_DRAIN_READY_LABEL_KEY | Label key marking a node as drain ready. Set only if the chart value is not empty. | "kaas.mirantis.com/csi-drained" |
cephDeployment.drainReadyLabelKey |
| DEPLOYMENT_CSI_DRIVERS_MANAGE | Enable CSI resources management by the Pelagia controller, such as CSI drivers or OperatorConfig. If disabled, other DEPLOYMENT_CSI_* options are not set and have no effect. |
"true" |
cephDeployment.csi.manage |
| DEPLOYMENT_CSI_RBD_DEFAULT_DRIVER_CREATE | Create a default CSI driver for RBD if not specified in the CephDeployment spec. |
"true" |
cephDeployment.csi.defaultDriversCreate.rbd |
| DEPLOYMENT_CSI_CEPHFS_DEFAULT_DRIVER_CREATE | Create a default CSI driver for CephFS if not specified in the CephDeployment spec. |
"true" |
cephDeployment.csi.defaultDriversCreate.cephfs |
| DEPLOYMENT_CSI_NFS_DEFAULT_DRIVER_CREATE | Create a default CSI driver for NFS if not specified in the CephDeployment spec. |
"false" |
cephDeployment.csi.defaultDriversCreate.nfs |
| DEPLOYMENT_CSI_KEEP_EXISTING_ON_UPGRADE | When enabled, Pelagia does not change existing CSI drivers or OperatorConfig resources that it did not create. | "true" |
cephDeployment.csi.keepExisting |
| DEPLOYMENT_CSI_KUBELET_PATH | Path to kubelet on a host. | "/var/lib/kubelet" |
cephDeployment.csi.kubeletPath |
| DEPLOYMENT_CSI_ENABLE_CSIADDONS | Deploy CSI addons. | "false" |
cephDeployment.csi.addons |
| DEPLOYMENT_CSI_CONTROLLER_PLUGIN_NODEAFFINITY | Node affinity settings for the CSI Controller plugin deployment (for example, provisioner). | "" |
cephDeployment.csi.placement.nodeAffinity.controllerPlugin |
| DEPLOYMENT_CSI_CONTROLLER_PLUGIN_TOLERATIONS | Toleration settings for the CSI Controller plugin deployment (for example, provisioner). | "" |
cephDeployment.csi.placement.tolerations.controllerPlugin |
| DEPLOYMENT_CSI_NODE_PLUGIN_NODEAFFINITY | Node affinity settings for CSI plugins. | "" |
cephDeployment.csi.placement.nodeAffinity.nodePlugin |
| DEPLOYMENT_CSI_NODE_PLUGIN_TOLERATIONS | Toleration settings for CSI plugins. | "" |
cephDeployment.csi.placement.tolerations.nodePlugin |
| DEPLOYMENT_MANAGE_VOLUMESNAPSHOTCLASSES | Create default VolumeSnapshotClasses for Ceph CSI RBD and CephFS drivers. Requires the volumesnapshotclasses.snapshot.storage.k8s.io CRD. |
"true" |
cephDeployment.manageVolumeSnapshotClasses |
The DEPLOYMENT_CEPH_IMAGE and DEPLOYMENT_ROOK_IMAGE options are derived from the values of the images section.
For details, see Configuration example for Ceph and Rook images.
Options configurable manually#
The Pelagia Helm chart update does not affect the following Pelagia controller options.
Therefore, you can update them only manually using the kubectl -n pelagia edit cm pelagia-lcmconfig command.
| Parameter | Description | Default |
|---|---|---|
| DEPLOYMENT_LOG_LEVEL | Log level of the Pelagia deployment controller. Possible values (case-insensitive): trace, debug, info, warn, error. |
"info" |
| HEALTH_CHECKS_CEPH_ISSUES_TO_IGNORE | Comma-separated list of Ceph cluster health issues to ignore in the health state. Deprecated in favor of spec.cluster.healthCheck.muteHealthWarning. For details, see Rook documentation: healthCheck.muteHealthWarning. |
"OSDMAP_FLAGS,TOO_FEW_PGS,SLOW_OPS,OLD_CRUSH_TUNABLES,OLD_CRUSH_STRAW_CALC_VERSION,POOL_APP_NOT_ENABLED,MON_DISK_LOW,RECENT_CRASH" |
| HEALTH_CHECKS_SKIP | Comma-separated list of checks to skip during Ceph cluster verification. Possible values: ceph_daemons, ceph_csi_daemons, usage_details, ceph_events, pools_replicas, rgw_info, spec_analysis. |
"" |
| HEALTH_CHECKS_USAGE_CLASS_FILTER | Regexp-based filter to prepare usage details only for the specified device class. | "" |
| HEALTH_CHECKS_USAGE_POOLS_FILTER | Regexp-based filter to prepare usage details only for the specified pools. | "" |
| HEALTH_LOG_LEVEL | Log level of the Pelagia LCM health controller. Possible values (case-insensitive): trace, debug, info, warn, error. |
"info" |
| TASK_LOG_LEVEL | Log level of the Pelagia LCM osdremove-task controller. Possible values (case-insensitive): trace, debug, info, warn, error. |
"info" |
| TASK_OSD_PG_REBALANCE_TIMEOUT_MIN | Timeout in minutes to wait for an OSD to finish rebalancing to 0 before considering the rebalance failed. For the procedure, refer to CephOsdRemoveTask failure with a timeout during rebalance | "30" |
| TASK_ALLOW_REMOVE_MANUALLY_CREATED_LVMS | Remove LVM partitions during OSD partition cleanup, even if they were created manually. | "false" |