Backups & Storage
NAS storage and backup planner
Plan NAS usable capacity, parity vs cache behavior, daily-change headroom, and offsite copy gaps before buying drives or restructuring an array.
Use this before buying drives, adding parity, expanding a NAS, or trusting a NAS as the only backup target. Works for Unraid, Synology, TrueNAS, QNAP, and generic NAS.
Close the offsite copy gap before expanding storage.
An offsite copy is missing or unknown. The NAS plan is local-only until that gap is closed.
First checks
- Current data: 10 TB. Two-year projection (current + 2 * annual growth): 14 TB.
- Estimated usable capacity (conservative): 43.2 TB. Headroom after projection: 29.2 TB.
- Retention impact: 1500 GB across 30 days at 50 GB/day.
- Cache: 1 TB pool, approximately 10 days at current change rate before mover/flush.
Step-by-step troubleshooting
- 1
Verify current array health first
Open the platform's storage page (Unraid Main, Synology Storage Manager, TrueNAS Pools, QNAP Storage & Snapshots) and confirm no disk shows SMART warnings or degraded status.
Expected: Every existing disk is green/healthy and parity/scrub status is current.
Next: If any disk warns, follow the SMART triage page before expanding capacity.
- 2
Estimate usable capacity with vendor calculator
Use Synology's RAID calculator, Unraid's parity guidance, or TrueNAS pool math to confirm usable capacity for the planned layout. This tool is conservative; the calculator is exact.
Expected: Vendor calculator returns a usable TB number within ~10% of this estimate.
Next: If the vendor number is materially different, trust the vendor for purchase decisions.
- 3
Plan retention against usable capacity
Retention math: 50 GB/day * 30 days = 1500 GB. Compare against usable capacity and target headroom (>= 20%).
Expected: Retention fits inside usable capacity with headroom for growth.
Next: Retention is within range; revisit if daily change rate grows.
- 4
Close the offsite gap
Add an offsite copy (reputable cloud backup with versioning, or rotated USB kept off-site) with separate credentials from the local copy.
Expected: One copy survives loss of the home, with different credentials than the local copy.
Next: Run a restore drill from the offsite copy before deleting any source data.
- 5
Add power protection before scheduled writes
Confirm UPS battery age and shutdown integration are still configured correctly.
Expected: An outage triggers safe shutdown without interrupting scheduled writes mid-flight.
Next: Record UPS load watts and battery replacement date in the inventory.
What your answers suggest
- Raw capacity = 48 TB; parity-loss drives = 0; conservative usable (~90% of raw - parity) = 43.2 TB.
- Headroom after two-year projection (14 TB) = 29.2 TB.
- Retention impact = 1500 GB (3% of usable).
- Cache headroom = ~10 days at 50 GB/day before mover/flush.
Likely cause area
- Unraid: parity disk is separate from data disks; cache pool is independent of the array.
- Single parity protects against one data-disk failure; not a backup.
- No offsite copy reported — a fire, theft, or ransomware event takes every copy on the NAS.
- Capacity covers two-year projection at the current growth rate.
- Cache headroom is adequate for the current change rate.
- UPS protection reported or workload is lower-risk.
Safe actions
- Parity is not a backup — it protects against one (or two) disk failures, not ransomware, accidental delete, fire, or theft.
- Verify exact usable capacity with the platform's official RAID/storage calculator before buying drives.
- Add an offsite copy (cloud backup with versioning OR rotated USB) with credentials separate from the local copy.
- Run a non-correcting parity/scrub check after adding any new disk; do not start with corrections enabled.
- Keep UPS battery age recorded and replace before five years.
When to stop
- Stop before buying drives if current parity/scrub status is unknown or any disk is showing SMART warnings.
- Stop before expanding an array that has no current backup of its irreplaceable shares.
- Stop before rebuilding with mismatched cable/power gear that has not been re-seated.
- Stop before treating RAID/parity as a backup substitute.
Assumptions
- Conservative usable capacity assumes ~10% filesystem and reserved overhead; exact figures depend on platform, filesystem, and reserved metadata.
- Parity loss assumes user input matches the convention for the selected platform (Unraid parity disks are extra; RAIDZ/mirror/SHR parity comes out of the pool).
- Retention math is linear; deduplication, compression, and snapshot reflinks can reduce actual storage materially.
- Cache day estimate uses 50% of cache capacity as a safe operating headroom; sustained 100% utilization is not safe to plan around.
What should I check first?
- Confirm current data size, expected growth, and the daily change rate that backups must absorb.
- Check the array's parity/RAID mode and whether a cache pool exists separately from the array.
- Confirm whether any current data already has an offsite copy that survives loss of the home.
What is likely wrong?
- Parity is being treated as a backup instead of an uptime feature.
- The cache pool is undersized for the daily change rate, so writes spill to the array and slow everything down.
- Retention math has not been done — "keep forever" plus a high daily change rate runs out of space quickly.
- Offsite copy is missing, weak, or shares an account with the local copy.
What is safe to try?
- Use conservative estimates: usable capacity below the marketing number, runtime below the marketing number.
- Plan one local copy plus one offsite copy with separate credentials before adding more bays.
- Add UPS protection before relying on overnight backups.
When should I stop?
- Stop before buying drives if current parity status is unknown or any disk is showing SMART warnings.
- Stop before expanding an array that has no current backup of its irreplaceable shares.
- Stop before rebuilding a parity disk with mismatched cable/power gear that has not been re-seated.
Source-backed checks
HomeTechOps turns official docs and conservative safety rules into a shorter runbook. These links are the source trail for the page direction.