RAID Rebuild Time Estimator
Estimate how long an array runs degraded after a drive fails.
What you provide
A 7,200 rpm drive sustains roughly 1,200 Mbps sequentially. Controllers usually rebuild slower.
Result
CHECKAbout 127.0 hours degraded on a RAID 5 rebuild
For 127 hours this array survives no further failures. Drives bought together tend to fail together, they are the same age and have done the same work, and a rebuild puts every one of them under sustained load. Two practical mitigations: lower the rebuild priority only if you can accept the longer window, and make sure a genuine backup exists, because a rebuild is a repair and not a copy.
- Array
- RAID 5, 5 × 8 TB
- Data read to rebuildParity reads every surviving member
- 32.0 TB
- Idle array
- 88.9 h
- Serving 30% load
- 127.0 h
- Degraded window
- high
- Second failure toleratedAnother failure during rebuild loses the array
- no
What this cannot tell you
- An estimate from the figures you enter. It cannot see your array, your controller or your drives.
- Real rebuild times vary enormously with controller behaviour, rebuild priority settings and how full the array is. Treat this as an order of magnitude, not a schedule.
Take this with you
How this calculation works
A mirror only has to copy one drive, while a parity level has to read every surviving member to reconstruct the missing one. That difference is why rebuild time grows with array width as well as with drive size, and why a wide RAID 5 of large drives spends so long exposed. The estimate divides the data that must be read by the rebuild throughput, then stretches it to account for the array still serving clients, since almost no rebuild happens on an idle array.
What the results mean
- Data read to rebuild
- Everything the controller has to read to reconstruct one drive. On parity levels this is the whole array minus the parity drives, not just one disk.
- Degraded window
- How long the array survives no further failures. This is the actual risk being measured, not the inconvenience of a slow rebuild.
- Second failure tolerated
- Whether another drive can die mid-rebuild without losing data. RAID 6 says yes. Everything else says no.
Common problems and fixes
- The rebuild is far slower than the estimate
- Check the rebuild priority setting on the controller, which usually defaults to favouring client traffic over rebuild speed. Also check whether the array is being scrubbed at the same time, and whether one of the surviving drives is itself struggling. A drive that is slow but not yet failed will drag a rebuild out for days and rarely announces itself.
- A second drive failed during the rebuild
- This is common enough to be expected rather than unlucky. Drives bought together are the same age and have done the same work, and a rebuild is the heaviest sustained read the array ever performs, so it finds the drive that was quietly failing next. Stop writing to the array and restore from backup rather than forcing the array back online, which can make the damage worse.
Frequently asked questions
Why does RAID 5 take longer to rebuild than RAID 1?
A mirror copies one drive to another. Parity has to read every remaining drive in full and compute the missing data from all of it, so an eight-drive RAID 5 of 8 TB disks reads 56 TB to rebuild one. Widening a parity array increases capacity efficiency and lengthens the window where it is vulnerable.
Should I use RAID 6 instead of RAID 5?
On large drives, yes. The reason is arithmetic rather than taste: a RAID 5 rebuild has to read tens of terabytes without a single unrecoverable error, and at the consumer drive specification of one error per 10^14 bits, the odds stop being comfortable. RAID 6 keeps a second parity drive so one read error during rebuild is survivable.
Is a rebuild a backup?
It is a repair, not a copy. A rebuild restores redundancy that was lost, and it does nothing at all for a file you deleted, a database that corrupted itself or a ransomware run that encrypted the volume. Any array, at any level, still needs backups that live somewhere it cannot write to.
Put this on your own site
Free to embed, no attribution required beyond the source link the frame carries itself. It runs entirely in your visitor's browser, sets no cookies and loads no third-party script.
<iframe src="https://runthetests.com/embed/raid-rebuild-time/" width="100%" height="560" style="border:1px solid #e5e5e5;border-radius:8px" title="RAID Rebuild Time Estimator" loading="lazy"></iframe>
Preview it at https://runthetests.com/embed/raid-rebuild-time/. Embedded pages are marked noindex, so yours stays the canonical copy — not this one.