RAID Rebuild Time: How Long Does It Actually Take?
A drive failed, you've installed the new one, the rebuild is running - and you wonder: how much longer? Here's the honest math with an interactive calculator and real numbers.
Short version
A 16 TB drive typically takes 12-24 hours to rebuild at 100-180 MB/s sequential write speed. With many drives in RAID 5/6 it can rise to 30-50 hours because parallel reads from all drives become the bottleneck. During rebuild the array is unprotected (RAID 5) or running at reduced redundancy (RAID 6).
Rebuild Time Calculator
What affects rebuild time
Drive size is the main factor. Linear scaling: 4 TB takes half as long as 8 TB.
Write speed of the new drive. Modern 16-22 TB drives push 200-280 MB/s sequentially. On classic 7200 RPM CMR drives realistically 150-180 MB/s during rebuild.
Number of drives matters a lot for RAID 5/6/Z. Rebuild reads from all remaining drives. More drives = more parallel reads = often controller or PSU bottleneck.
RAID type. RAID 10 rebuilds only from the mirror partner = fast. RAID 6 has to compute two parities = slightly slower than RAID 5.
Active load. If the NAS is in normal use during rebuild, that drops rebuild speed to 50-70 percent.
Understanding URE risk
During rebuild every remaining drive is read in full. At consumer URE rate of 1 in 10^14 bits, a rebuild of 5x16 TB (= 80 TB = 6.4x10^14 bits) statistically encounters ~6.4x the URE limit.
RAID 5 isn't protected against a single URE during rebuild - rebuild aborts. RAID 6, RAIDZ2, and Unraid 2P can absorb it.
More: RAID 5 vs RAID 6.
How to speed up rebuild
Synology / QNAP: Set rebuild priority to "fast" (Storage Manager). Trade-off: user-visible performance drops.
mdadm (Linux): echo 200000 > /proc/sys/dev/raid/speed_limit_max raises the cap. Default is often conservative.
ZFS: zpool scrub plus resilver performance via vfs.zfs.resilver_min_time_ms, but be careful.
Pause active workloads. Stop Plex streams, suspend backup jobs. Often halves the time.
What to do during rebuild
1. Verify your backup. Confirm the last successful backup. If a second drive fails during rebuild, that's your insurance.
2. SMART check all remaining drives. Reallocated sectors, pending sectors on every other drive. If one shows signs: replace after rebuild too.
3. Leave the NAS alone during rebuild. No big write jobs, no reboots, no maintenance.
4. Guard against power loss. A UPS is mandatory anyway (UPS guide) - critical during rebuild.
If rebuild aborts
RAID 5: often data loss without backup. Pro recovery is possible but expensive (1500-5000 USD).
RAID 6/RAIDZ2: second parity catches URE. Rebuild continues despite individual read errors.
Prevention: from 12+ TB drives or 5+ bays go RAID 6 or RAIDZ2. More in the RAID guide.
Recommendation
Plan for rebuild time before you need it:
- 4-6 drives and drives up to 8 TB: RAID 5 acceptable
- Larger drives or more bays: RAID 6, RAIDZ2 or Unraid with 2 parity
- Plan a hot spare (see Hot vs Cold Spare) so rebuild starts immediately
- NAS on UPS, regular scrubbing as a pre-check
Related articles
Further reading
RAID 0 Explained: Speed Without a Safety Net
RAID 5 vs RAID 6: Which Should You Actually Pick?