FEAT-080 - Refresh Protected-Volume Disk-Space Snapshots In The Background¶

Summary¶

The download queue periodically checks protected temp, incoming, and config volumes so it can pause downloads before disk pressure corrupts state. Current main has already improved this path with cached volume identities and protected volume snapshots, but the refresh still runs from the download-queue processing path.

Move expensive or unreliable filesystem/volume probes behind a background snapshot refresher while keeping the download queue's stop/pause decisions on the existing owner path.

Intended Shape¶

Maintain a last-known protected-volume snapshot refreshed by a bounded worker.
Let the download queue consume a recent snapshot during its normal 100 ms processing loop.
Fall back to the current synchronous behavior when no valid snapshot exists.
Treat unresolved or stale protected-volume data conservatively; never weaken the disk-space guard.
Keep explicit diagnostics for stale snapshots, unresolved volumes, and worker refresh failures.

Scope Constraints¶

Do not move download state mutation to the worker.
Do not weaken disk-space fail-closed behavior.
Do not make network/removable volume stalls block UI or queue processing when a previous safe snapshot is available.
Do not combine this with temp-directory placement policy changes.

Acceptance Criteria¶

[ ] protected-volume snapshots can refresh in a background worker
[ ] download queue consumes snapshots without mutating state from the worker
[ ] stale, missing, or unresolved snapshots remain fail-closed
[ ] diagnostics distinguish synchronous fallback, worker success, stale data, and worker failure
[ ] tests cover stale snapshot policy, unresolved volume policy, and breach action consistency
[ ] live profile with slow/removable paths does not stall queue processing

Validation¶

python -m emule_workspace validate
focused native disk-space guard seam tests
live or synthetic slow-volume probe test
x64 Debug and Release app builds before commit