Network & Infrastructure / Storage Networks & Distributed Data
Device-class rules and capacity stranded in another tier
A fast pool cannot rebalance despite free disks elsewhere in the cluster. Its rule may intentionally limit placement to a device class whose own capacity is exhausted.
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
Compare pool CRUSH rules, class assignments, eligible OSD count and utilization within that class. Check whether replacement disks were automatically classified as expected.
Explain the result
Device classes constrain eligible storage, so cluster-wide free bytes do not guarantee usable capacity for every pool. Mixing classes can also violate the workload's performance assumptions.
Plan the fix
Add or rebalance capacity within the approved class, or plan an explicit pool policy change with performance and resilience review. Do not relabel devices simply to silence a capacity warning.
Check the result and rollback
Validate placement, client latency and recovery headroom after the change. Restore the previous rule if data lands on an unintended tier or the eligible set cannot satisfy failure-domain requirements.