Ceph Status Details, OSDs and Placement Groups

The ClusterCeph Status areaand exposesObjects / PGs tabs expose detailed Ceph statushealth information that is useful for operations and troubleshooting. The live eEKAS R8 interface includes tabs for Ceph Status, Monitors, Managers, MDS, Objects / PGs and MON/MGR Placement.

Key Health Fields

Field	Meaning
Health Status	Overall Ceph health state. HEALTH_OKHealthy means no critical health issue is currently reported.
ReasonTotal MONs	Short explanationNumber of monitor daemons known to the current health state.cluster.
OSDsActive up/in/totalMGR	ShowsThe howmanager manydaemon OSDscurrently are running, part of the cluster and known in total.active.
RawActive UsedMDS	PercentageThe ofactive rawmetadata clusterserver, capacityif currentlyCephFS/MDS services are used.
PG PressureStates	Placement group pressurestate indicator. The dashboard can show maximum and average values.
Failure Domain	The failure boundary used for data placement, for example host.summary.

OSD Interpretation

OSDs areIn the storagetest daemonscluster, thatCeph storereported object data. Aa healthy productionstate, stateTotal normallyMONs requireswas 5, the expectedactive OSDsmanager towas be both upEEKASBG-1.ynuwsd and in.PG Ifstates anshowed OSDactive+clean: is180.

down or out, the cluster may still serve data, but redundancy

Objects and recovery state must be checked before additional changes are made.

Placement Groups

Placement

groups distribute data across OSDs.

The Objects / PGs viewtab helpshighlights administratorsobject-level understandrisk objectindicators.

distribution, Highoruneven PG pressure can indicate that capacity, placement or recovery behaviour should be reviewed.

Field	Healthy Meaning
Degraded Objects	0 means no objects currently lack the expected redundancy.
Misplaced Objects	0 means data placement currently matches the intended CRUSH placement.
Unfound Objects	0 means no objects are missing from all known OSDs.
PG countsStates	active+clean means placement groups are active and pressure.clean.

Recommended Troubleshooting Flow

1. Start with the dashboardCluster healthStatus summary.tab.
2. Open ClusterCeph Status and reviewcheck CephHealth Status.Status, Total MONs, Active MGR and PG States.
3. Check OSDs up/in/total and raw usage.
4. ReviewOpen Objects / PGs whenand performance,check balancingdegraded, ormisplaced recoveryand behaviourunfound looks unusual.objects.
5. Record the health state and the exact warning textvalues before taking corrective action.

Good Practice

Do not perform

Avoid large configuration changes duringwhile andegraded, unhealthymisplaced stateor unfound objects are present unless the change is part of the recovery plan.

First document

Good Practice

A healthy state should show zero degraded, misplaced and unfound objects. Any non-zero value should be investigated before handing the currentservice status,to then decide whether the issue is node, disk, network, capacityusers or placement related.customers.