Calculate RAID rebuild time from drive capacity, speed, RAID level, and load, plus URE risk and the drive speed needed to finish within a target window.
RAID Rebuild Time Formula
T = (C * 1,000,000 * U) / (S * R * L * 3600)
- T is the rebuild time in hours
- C is the failed drive capacity in TB (decimal, 1 TB = 1,000,000 MB)
- U is the fraction of the array that must be rebuilt (1.0 for a full-drive hardware RAID rebuild, or percent used / 100 for a ZFS resilver or Synology fast repair)
- S is the per-drive sequential read speed in MB/s
- R is the RAID overhead factor (0.9 for RAID 1/10, 0.7 for RAID 5/50, 0.55 for RAID 6/60)
- L is the load factor (1.0 idle, 0.85 light, 0.65 moderate, 0.4 heavy production load)
The calculator’s default mode evaluates this formula directly: it converts the failed drive’s capacity into megabytes, multiplies the base drive speed by the RAID and load factors to get an effective rebuild speed, and divides data by speed to get the time the array will spend degraded. The required-speed mode rearranges the same formula to solve for S, which tells you how fast each drive must read to finish the rebuild inside a maintenance window you choose:
S = (C * 1,000,000 * U) / (T * R * L * 3600)
The advanced option estimates the probability of hitting an unrecoverable read error (URE) while reading the surviving drives during the rebuild:
P = 1 - e^(-(N * C * 8 * 10^12) / E)
- P is the probability of at least one URE during the rebuild
- N is the number of surviving drives that must be read in full (drives minus 1 for parity RAID, 1 for a mirror)
- C * 8 * 10^12 is the bits read per drive (C in TB)
- E is the drive’s URE rating in bits (10^14 consumer HDD, 10^15 enterprise HDD, about 10^16 for SSDs)
Typical Rebuild Times and Priority Settings
The table below shows estimated rebuild times for a full-drive rebuild using 7200 RPM hard drives (160 MB/s sequential) under moderate production load. Halve these numbers roughly for an idle array, and scale them down proportionally for a ZFS resilver on a partly filled pool.
| Failed Drive Size | RAID 10 | RAID 5 | RAID 6 |
|---|---|---|---|
| 4 TB | 11.9 h | 15.3 h | 19.4 h |
| 8 TB | 23.7 h | 30.5 h | 38.9 h |
| 12 TB | 35.6 h | 45.8 h | 58.3 h |
| 16 TB | 47.5 h | 61.1 h | 77.7 h |
| 20 TB | 59.4 h | 76.3 h | 97.1 h |
| 24 TB | 71.2 h | 91.6 h | 116.6 h |
Rebuild speed is usually governed by a priority or throttle setting, not by raw drive speed. If your rebuild is crawling, this setting is the first thing to check. The table below shows where that control lives on common platforms and what it does to the numbers above.
| Platform | Setting | Default | Effect on Rebuild Time |
|---|---|---|---|
| Linux mdadm | dev.raid.speed_limit_min / speed_limit_max sysctls | 1,000 / 200,000 KB/s | Raising speed_limit_min forces the resync to keep pace even under load |
| Synology DSM | Storage Manager resync speed limits (Faster Resync mode) | Balanced (throttled) | Faster Resync can cut rebuild time by half or more; SHR fast repair also skips unused blocks |
| QNAP QTS | Storage and Snapshots, RAID resync priority | Default (middle) | Resync First maximizes rebuild speed; Service First protects users and stretches the rebuild |
| TrueNAS / ZFS | Resilver Priority schedule | Off (throttled resilver) | Enables full-speed resilver inside a time window you define; resilver already copies only used data |
| Dell PERC / Broadcom MegaRAID | Rebuild Rate (0 to 100) | 30 | At 30 the controller reserves roughly a third of I/O for the rebuild; raising it toward 100 speeds the rebuild at the cost of host performance |
RAID Rebuild Time Example Problems
Example 1: How long to rebuild an 8 TB drive in RAID 5. A file server runs RAID 5 on 7200 RPM drives that sustain 160 MB/s, and it stays in production during the rebuild under moderate load. The effective speed is 160 * 0.7 * 0.65 = 72.8 MB/s. The data to rebuild is 8 * 1,000,000 = 8,000,000 MB. Time is 8,000,000 / 72.8 = 109,890 seconds, which is 30.5 hours, or about 1 day 6 hours 32 minutes. The array pace works out to about 262 GB per hour.
Example 2: What speed is needed to rebuild 16 TB in 24 hours. You want a 16 TB drive in a RAID 6 array rebuilt within a 24 hour window while light production load continues. The required effective speed is 16,000,000 / (24 * 3600) = 185.2 MB/s. Dividing by the RAID 6 factor (0.55) and the light load factor (0.85) gives a required base drive speed of 185.2 / 0.4675 = 396.1 MB/s. That is beyond any hard drive’s sustained rate, so you would need SATA SSD class drives, a longer window, or an idle array to hit the target.
RAID Rebuild Time Frequently Asked Questions
Why does my controller estimate a longer time than this calculator? The calculator assumes clean sequential reads at a steady effective speed. Real rebuilds slow down for random production I/O, controller throttling (many default to reserving only about 30% of bandwidth for the rebuild), bad sector retries on aging surviving drives, and thermal limits in dense enclosures. Treat the calculator’s result as a best-case floor, and if your controller reports several times longer, raise the rebuild priority or schedule the rebuild for off-peak hours.
Is RAID 5 still safe with large drives? The risk is less about time and more about what happens while reading. Rebuilding one 8 TB drive in an 8-drive RAID 5 array reads 56 TB from the surviving disks. At an enterprise URE rating of 1 per 10^15 bits, the chance of hitting an unrecoverable read error during that rebuild is about 36%, and with consumer drives rated 1 per 10^14 it approaches 99%. Since a degraded RAID 5 array has no redundancy left, a URE hit means lost data on that stripe and often a failed rebuild. This is why RAID 6 or RAID 10 is generally recommended once drives exceed about 4 TB.
Does a hot spare make the rebuild itself faster? No. A hot spare does not change rebuild speed, but it removes the delay between failure and rebuild start, since the array begins rebuilding automatically instead of waiting for someone to swap a drive. Total degraded time is the wait time plus the rebuild time, so a hot spare can still cut hours or days off your real exposure window even though the rebuild portion is unchanged.
