Network & Infrastructure / Storage Networks & Distributed Data
Near-full OSDs despite comfortable aggregate free space
A cluster reports substantial total free capacity while individual OSDs approach fullness thresholds. Placement constraints can leave free space unusable for the affected data.
Environment and inputs: Ceph and storage transport dependencies; match health identifiers and recovery behavior to the installed release. Diagnostic/design example; not tested on production equipment.
Collect the evidence
Inspect per-OSD utilization, pool rules, device classes, CRUSH placement and recovery headroom. Compare logical free space with where replicas or erasure-coded shards are actually allowed to land.
Explain the result
Aggregate capacity hides distribution and placement constraints. A full device in a constrained failure domain can block work even when another class or location has abundant space.
Plan the fix
Correct the constrained capacity or placement design through the supported rebalance process. Avoid raising fullness limits as the primary remedy without a reviewed emergency capacity plan.
Check the result and rollback
Track the fullest devices and projected recovery placement, not only cluster averages. Revert placement changes if they violate failure-domain requirements or move excessive data through an already stressed cluster.