Short answer: You expand a NAS with bigger drives by replacing them one at a time — always a single drive, followed by a complete rebuild, and only once that rebuild is 100 % finished do you move on to the next one. Every replacement is the same stress test for the remaining drives as repairing a degraded array: the controller has to read a huge portion of the surface of every member, and that is exactly when latent bad sectors surface — the ones normal day-to-day use never touched for years. With classic RAID 5/6 the new capacity only appears after all drives have been swapped; with Synology SHR it usually shows up once two bigger drives are in place. Before you pull the first drive, verify your off-NAS backup by actually opening files from it and check the SMART data of every drive — if the array reports degraded status or any drive shows bad sectors, postpone the expansion.
A NAS bought five years ago with what were then generous drives tends to be full today. Photos, computer backups, camera footage, and in companies project data and shared folders — and the admin interface has gone orange on capacity. There is a temptingly simple solution: buy bigger drives and swap them in. But expanding a live array is technically the same thing as a series of rebuilds back to back, and a rebuild is the single operation that most often sends collapsed arrays to our lab. This guide covers how to do the expansion in a sensible order, what happens under the hood, and exactly where the risk of data loss lies.
Three ways to increase NAS capacity
Before you start ordering drives, get clear on which option you actually need.
1. Adding a drive to a free bay. If the NAS has an empty slot, expanding the array online with another member is usually the cheapest route. Even here, though, data is recalculated and redistributed across all drives — in terms of load it is comparable to a rebuild, just with a different outcome.
2. Replacing drives one by one with bigger ones. The most common scenario for fully populated NAS units with two to four bays. Drives are swapped one at a time, each replacement triggers a rebuild, and the volume is expanded at the end. This is the route the rest of the article is about.
3. A new array and data migration. The safest option: build a new NAS or a new array from new drives, copy the data across, and leave the old array running until the new one is verified. It does require a second device or enough bays, so most households and small businesses skip it — and take on a live swap instead.
If you are not sure which route makes sense for your device and your data, it is cheaper to get advice before buying drives than after a rebuild has stalled. You can arrange a consultation on NAS units and disk arrays with us through our contact page.
SHR vs. classic RAID: when the new capacity really appears
This is where most of the disappointment comes from, so let’s be blunt.
Classic RAID 5/6 (Synology in RAID mode, QNAP, TrueNAS, most off-the-shelf NAS units) calculates usable capacity from the smallest member of the array. If you swap one drive for an 8 TB unit in an array of 4× 4 TB, you gain nothing — half of the new drive sits idle until you replace all four and run the volume expansion. And the whole time you are paying the full price in risk: four replacements mean four complete rebuilds.
Synology SHR (Synology Hybrid RAID) builds the array from several layers and can combine drives of different sizes. The extra capacity of the bigger drives becomes available as soon as there are at least two larger units in the array — because that added portion needs redundancy too. So in a four-bay SHR setup you will see the first gain after the second replacement, not the fourth. SHR-2 behaves the same way, with a tolerance of two drives.
QNAP offers online replacement and expansion for classic RAID („Replace & Expand"), but the smallest-drive rule applies just the same: capacity only jumps once every member has been swapped.
The practical upshot: with classic RAID, consider whether route number 3 — a new array and migration — makes more sense than four consecutive rebuilds. With SHR, replacing drives one at a time is at least partly rewarded along the way.
Why expansion is as risky as a rebuild
A rebuild is not copying one drive. During it, the controller reads the entire relevant surface of all remaining drives, recalculates parity and writes it to the new member — hours to days of continuous full load. On older drives it is exactly this load that brings latent bad sectors, timeouts and weak spots to light, the ones normal operation never read. We covered why that reliably finishes an array off, using a real case, in the article on why a rebuild kills RAID arrays.
Compared with an ordinary replacement of a failed drive, expansion has three aggravating factors:
- You are not doing one rebuild, but a series. A four-bay NAS = four full rebuilds within a few days or weeks. The window of risk is multiplied.
- The drives in the array tend to be the same batch, the same age, the same hours. They were bought together and wore out together. When you replace the first one, the others are not in the condition they appear to be — and you are sending them straight into the toughest test of their lives.
- You are voluntarily pulling a healthy drive. With RAID 5 and SHR-1 this puts the array into a degraded state with zero tolerance for another error. A single unreadable sector on any of the remaining drives can bring the rebuild down.
And one more trap that gets worse with every drive generation: the bigger the new drives, the longer each rebuild takes and the longer the array runs without redundancy. A NAS bought five years ago is typically being expanded today from drives of a few terabytes to twenty-something — a rebuild of such a member has to move several times more data, yet the transfer speed of spinning drives has not grown anywhere near as much in the meantime. What used to take hours on a small array takes days at today’s capacities. That is exactly why double parity — RAID 6 or SHR-2 — is increasingly recommended for high-capacity arrays; it survives even an error on a second drive during a rebuild. If you go with route number 3 and rebuild the array from scratch, that is the only convenient moment to decide on the array type.
What to do before the first replacement
The order of the steps matters more than the brand of the drives.
- Verify your backup. For real. Not „the backup job is green", but opening a random selection of the files whose loss would hurt most. RAID is not a backup — expansion is precisely the situation where that becomes clear. How to put together a sensible copy scheme is described in the article on the 3-2-1 rule and an external drive as a backup.
- Check the SMART data of every drive, not just the capacity. Reallocated and pending sectors on any member mean stop — such a drive most likely will not survive a rebuild, and the expansion needs to be postponed, or that drive’s condition dealt with first.
- Run a data scrub / consistency check and let it finish. Better to find an unreadable sector now, with full redundancy, than in the middle of a degraded state.
- Document the array. Photograph the bays, note the serial numbers and drive positions, the array type, the system version. If something goes wrong, this documentation makes recovery considerably cheaper.
- Update DSM/QTS beforehand, not mid-process — a firmware upgrade has no place in a half-finished expansion. With older models, also check whether they still receive updates: manufacturers gradually end support for older generations, and for a device that no longer gets security patches it is worth asking whether to invest in new large drives at all, or move the data elsewhere instead.
- Plan for a quiet period. Expanding a four-bay NAS is a matter of days to weeks of pure rebuild time. Don’t do it during a deadline week, and don’t leave unnecessary load running on the NAS during the rebuilds.
Step-by-step replacement procedure
The specific buttons differ between Synology, QNAP and the rest; the logic is the same.
- Replace one drive. Either hot-swap it according to the manufacturer’s manual, or shut the NAS down properly, swap the drive and power it back on. Never pull two drives at once — with RAID 5/SHR-1 that is an immediate array failure.
- Start the array repair (Repair/Rebuild) in the admin interface over the new drive.
- Wait for 100 %. Let the rebuild finish completely, check that the array reports healthy status, and only then continue. No „it will catch up, I’ll swap the next one right away".
- Watch the logs and SMART during the rebuilds too. Read errors, timeouts or another drive flagged as faulty mean stop immediately — no repeated clicking on Repair.
- Repeat for the remaining drives, always with a completed rebuild between replacements.
- Expand the volume. After the last replacement (with SHR, after the second) DSM/QTS will offer to expand the storage pool and the volume; the file system stretches onto the new capacity.
- Don’t wipe the old drives right away. Set them aside, labelled and in their original order, until the expanded array has been running reliably for a few weeks and the backup passes its checks. They are your last safety net.
Which drives for the expanded array
Briefly, because we have a detailed comparison elsewhere:
- CMR only, no SMR. An SMR drive in a RAID means a rebuild that takes extremely long or never finishes at all. The difference between the technologies is explained in our CMR vs. SMR overview.
- NAS-grade drives, not desktop ones. Vibration in the array, continuous operation and behaviour on a read error (TLER/ERC) are exactly the parameters that decide whether a rebuild survives. Specific recommendations based on our lab experience are in the comparison of NAS drives WD Red, IronWolf and Toshiba N300.
- New drive capacity ≥ the original, and ideally all new drives the same size. Mixing brands is fine, mixing categories (NAS/CMR) is not.
- Check the NAS manufacturer’s compatibility list for your model — especially at high capacities. On newer devices this is not a formality: Synology has tightened drive compatibility on its latest Plus series, and some models require units from their own verified drive list when creating an array. So the rules differ by model and system version, and a drive that ran without trouble in an older NAS may not pass in a newer one.
When an expansion goes wrong
The typical scenarios that reach us from arrays being expanded: the rebuild stalls on read errors from another drive, the array drops into degraded status with no tolerance left, or „volume crashed" appears after a replacement and the volume won’t mount.
At that point there is only one safe rule: stop writing and stop repairing. Don’t click Repair over and over, don’t create a new array over the same drives, don’t change the bay order, don’t reset or reinstall the system. Each of these steps can overwrite the metadata without which the array can no longer be assembled. Power the array down and document the messages and drive positions — what the individual Synology messages actually mean and why an array must not be „treated" by clicking around in DSM is covered in the article on data recovery from a Synology NAS.
If the array holds data you cannot afford to lose, have its condition assessed before you start anything else. A simple initial assessment of RAID and NAS may be free (analysis of multiple drives or array reconstruction is quoted in advance) — you will find out what state the drives are really in and whether it is safer to finish the rebuild or to clone the drives and assemble the array offline. No-recovery-no-fee applies: if the requested data cannot be recovered, you pay nothing.
Frequently asked questions
Does capacity increase right after the first drive is replaced?
With classic RAID 5/6, no — the new capacity only becomes available once all drives have been replaced and the volume expanded. With Synology SHR the gain appears as soon as there are at least two bigger drives in the array.
Can I replace two drives at once to speed things up?
No. With RAID 5 and SHR-1 the array would lose all redundancy and a single read error would mean the volume goes down. Always one drive, a completed rebuild, then the next.
How long does a rebuild take after a drive replacement?
Hours to days, depending on capacity, how full the array is, the load and the condition of the drives. Don’t power the NAS off during rebuilds, don’t pile extra load on it, and start the next replacement only after the current rebuild is 100 % complete.
Do the new drives have to be the same brand as the originals?
No. What matters is that they are CMR drives from NAS ranges, with a capacity at least equal to the original members, ideally from the NAS manufacturer’s compatibility list.
What about the old drives after the replacement?
Set them aside, labelled and in their original order, until the expanded array has been running reliably for several weeks and the backup passes its checks. Until then they are the last copies of your data; only after that should you wipe them or use them elsewhere.
The rebuild during expansion failed and the volume is unavailable. What now?
Power the array down, don’t repair anything, don’t create a new array and don’t change the drive positions. Photograph the messages and the bay order and have the drives diagnosed — as part of data recovery a simple initial assessment may be free, more demanding analysis is quoted in advance. The chances of recovery are highest as long as nobody has overwritten the array metadata with further attempts.
Where to find us
ITHOPE is a company based in Brno, Czechia: we do data recovery from disks, RAID and NAS, PC and laptop service and IT for business, all in our own lab at Viniční 3587/216, 615 00 Brno.
Phone +420 775 556 063 (24/7 line). You can bring the device in person, or send it by courier (Zásilkovna, PPL).