Explain pending space after Proxmox Backup pruning

Distinguish snapshot removal from chunk reclamation when a Proxmox Backup datastore still looks full.

Source facts

Proxmox documents pruning as removing snapshot metadata rather than immediately deleting its data chunks. Other snapshots may still reference those chunks. Garbage collection identifies unused chunks through mark and sweep phases, with a cutoff affected by active backup writers and a grace period. Chunks within that period appear as pending removals. Consequently, deleting a snapshot and reclaiming its storage are different events.

Source: Proxmox — Maintenance Tasks, Backup Server documentation.

Applicability

Use this guide for the Proxmox Backup datastore behavior described in the maintenance documentation. Pruning alone is also not evidence that all underlying sensitive data has been removed.

DSE recommendation

DSE recommends reviewing retention intent before pruning additional recovery points to solve a capacity alert. Record datastore use, the last prune result, garbage-collection status, pending removals, and active writers together. Use the prune simulator when evaluating a retention change. Assign a capacity escalation owner so urgent space pressure does not become repeated deletion without a recovery decision. Keep filesystem-level chunk deletion outside the troubleshooting procedure and use the product’s supported maintenance operations.

Verification

Compare datastore use and garbage-collection reports across the relevant grace period, noting active jobs and references that remain. Confirm required recovery points are still present and follow the approved restore-test process. Close the capacity issue on measured reclaimed space and retained recovery coverage.

Official reference

Proxmox — Maintenance Tasks, Backup Server documentation. Reviewed October 14, 2026.