Parity Checks & Mover

Overview

UnRAID maintenance centers on two automatic background processes: parity checks (data integrity verification) and Mover (cache-to-array file migration). Neither is configured to run on a schedule in the current diagnostics snapshot — both should be reviewed.

Parity Checks

A parity check reads every data disk simultaneously, recomputes parity, and compares it against the parity drive — it confirms the array is internally consistent, not that the data itself is uncorrupted.
Non-correcting checks (report-only) are the safe default for routine scheduling; correcting checks should only run after an unclean shutdown or to fix errors a non-correcting check already found.
UnRAID automatically triggers a parity check after any unsafe shutdown — using a UPS is the primary defense against triggering this unnecessarily.
Recommended cadence for this array size (6TB drives): monthly, scheduled during low-usage overnight hours via Settings → Scheduler → Parity Check. A full check on a 6TB array typically runs 4–12 hours depending on concurrent load.

Mover

Mover transfers files between cache and array based on each share's Primary/Secondary storage and cache-use settings.
On this host, appdata/docker/system are cache-only (Mover takes no action on them) — Data/Photos/Multimedia write directly to the array (Mover also has nothing to move for these, since they bypass cache entirely).
Since none of the array-facing shares use cache as a write buffer, Mover schedule is largely inactive on this host today — worth confirming this is intentional rather than an oversight, since it means large incoming writes to those shares go straight to spinning disk at HDD write speed rather than bursting to SSD first.

Notes / Gotchas

Always disable Docker and VM Manager before manually running Mover — open file handles from running containers can cause incomplete or failed moves.
Mover activity is also the trigger for the NFS stale-file-handle issue documented on the Gotchas page — when Mover relocates a file, its underlying fileid changes, and NFS clients holding the old handle break.

Last Updated

2026-06-23