Your Expensive SSD Could Lose Data Without Warning

An SSD can stop working with no clicking, no slowdown and no warning, and when it does the data on it may be gone in one step. Speed and price do not change that: the only reliable protection is a tested backup kept somewhere else, with drive health monitoring and clean power to cut the odds of needing it.

This post explains how SSDs fail, what SMART monitoring can and cannot tell you, why power-loss protection matters on business drives, and how a 3-2-1 backup plan works in a small office. Products mentioned include the Micron 5400 PRO and Micron 7450 PRO enterprise SSDs, the Synology DiskStation DS925+ and QNAP TS-464 NAS, and the APC Back-UPS Pro BR900MI.

How do SSDs fail, and which failures give a warning?

The pattern is uncomfortable: only wear-out reliably announces itself. A hard disk often gets noisy or slow before it dies; an SSD has no moving parts to complain, so the first sign of a controller fault can be a missing drive letter on Monday morning. SSDs are not fragile; their failures just tend to be binary.

Is the SSD itself the problem?

To be fair to the hardware, a good SSD is far less likely to lose data than the habits around it. Micron rates its 5400 data center SATA SSD at an MTTF of 3 million device hours and backs it with a 5-year warranty. But MTTF is a statistic across a large population of drives, not a promise about the one in your server. A warranty replaces the drive; it does not return the files.

The other side of the argument is that business data now lives on SSDs by default, so even rare failures add up across a fleet. Our view: treat every SSD, however expensive, as a component that will fail eventually, and design so that the failure costs you a replacement part rather than your accounts or customer records.

An expensive SSD buys you speed and a better chance of survival. It does not buy you a backup.

What do TBW and DWPD actually mean?

Endurance is the one failure mode you can plan for. Drive makers quote it as TBW (terabytes written) or DWPD (drive writes per day over the warranty period). Micron's product brief rates the 5400 PRO at 1.5 DWPD for the 1.92TB model, and our listing for the Micron 5400 PRO 1.92TB SATA Enterprise SSD (S$3,881.00) gives 5,256 TBW with a 5-year warranty. The Micron 7450 PRO 960GB NVMe M.2 SSD (S$2,178.00) is listed at 1,700 TBW, also with a 5-year warranty, plus TCG Opal self-encryption.

Micron is careful to say these are theoretical maximums for a given transfer size and that "actual lifetime will vary by workload". A file server that is mostly read will rarely get near its rating; a database log or video cache can. If you want to understand why enterprise endurance costs more, our sibling post Why Enterprise SSDs Cost So Much More Than Consumer SSDs covers DWPD, latency and form factors in depth.

Can SMART monitoring predict SSD failure?

Partly. SMART (Self-Monitoring, Analysis and Reporting Technology) data from an SSD includes wear and error counters that are useful for spotting a drive approaching the end of its endurance or building up media errors. Business NAS systems and server management controllers can read these values and email an alert. Turn those alerts on and send them to a mailbox someone actually reads.

What SMART cannot do is predict a controller that dies outright or a capacitor that fails on a power cut. Treat a healthy SMART report as "no known problem", not "safe". Monitoring shortens the time between a problem and someone noticing; backup is what saves the data.

Why does power-loss protection matter?

SSDs hold data briefly in fast cache before committing it to flash. If power drops mid-write, data in flight can be lost and the drive's own mapping tables can be damaged. Micron describes the power-loss protection on its 5400 SSD as helping to "protect data in flight and data at rest from unexpected power loss", and our 7450 PRO listing notes full power-loss protection. Many consumer drives do not have this hardware, which is one reason we suggest enterprise SSDs for servers and NAS caches.

A UPS adds a second layer by keeping the system up through short dips and giving it time to shut down cleanly. The APC Back-UPS Pro BR900MI (S$452.50) is a 900 VA / 540 W line-interactive unit with AVR, suitable for a small NAS and network switch; its listing gives the VRLA battery an expected life of 3 to 5 years, so a UPS needs its own maintenance plan. Our sibling post Your UPS Battery Might Be Dead Before You Realise It explains how to check, and the UPS size calculator guide helps with sizing. If you want the NAS to shut itself down automatically on battery, confirm the UPS and NAS communication options with us before ordering.

What does a 3-2-1 backup look like in a small office?

The US Environmental Protection Agency's backup guidance tells organisations to "implement the NIST 3-2-1 rule": keep three copies (one primary and two backups), on two different media types, with one copy offsite. It also recommends isolating backups from network connections and regularly testing them, including the ability to recover individual files. Singapore's IMDA pre-approval guide for backup solutions asks for the same principles: copies kept separate and isolated from the operating environment, backup frequency set to business needs, and backups protected from unauthorised access.

A practical version for a 10 to 50 person office:

• Copy 1: the live data on laptops, servers or Microsoft 365.
• Copy 2: a NAS on site, with scheduled backups and snapshots. The Synology DiskStation DS925+ (S$1,260.00) has four bays, two M.2 NVMe slots and a 3-year Synology warranty on its listing; the QNAP TS-464 4-Bay NAS (S$1,080.00) has two 2.5GbE ports and a PCIe slot for a 10GbE card.
• Copy 3: offsite, in the cloud or on removable media such as LTO tape that is disconnected after each job. Our LTO-7 vs LTO-8 vs LTO-9 comparison covers tape.

Then test it. A restore drill every quarter, where someone recovers one file and one full folder, will find most backup problems before they matter. Isolation also matters for ransomware, which encrypts whatever it can reach; see preventing ransomware in Singapore.

Is RAID a backup?

No. RAID keeps a server or NAS running when one drive fails, which is valuable, but it copies deletions, corruption and ransomware to every drive in the array instantly. Use RAID for uptime and backups for recovery.

Buying SSDs, NAS and UPS in Singapore

SourceIT supplies Micron, Synology, QNAP and APC with local warranty and a GST invoice. Prices above are as listed on 23 Oct 2026, and every item mentioned was in stock at the time of writing. Browse Micron Enterprise and Data Center SSDs or Synology backup and data protection. For warranty terms and what a claim involves, read The Truth About Manufacturer Warranty. If you are weighing new versus second-hand storage servers, see Are Refurbished Enterprise Servers a Hidden Goldmine or a Huge Gamble? Tell us how much data you hold and how quickly you need it back, and we will propose a backup design: use Request for Quotation on the product page, call (65) 6978 3502, email Sales@sourceit.com.sg or visit our contact page.

Frequently asked questions

Can an SSD fail without any warning? Yes. Controller or firmware faults, component failures and power loss during writes can make an SSD unreadable with no prior symptoms. Wear-out is the failure mode most likely to show up early in SMART data.

What is power-loss protection on an SSD? Hardware on the drive that lets it finish committing data in flight when power is cut. Micron says the feature on its 5400 SSD helps protect data in flight and data at rest from unexpected power loss.

What does TBW mean on an SSD? Terabytes written: the total amount of data the maker rates the drive to absorb over its warranty. The Micron 5400 PRO 1.92TB is listed at 5,256 TBW. Actual life depends on workload.

Does SMART monitoring prevent data loss? No. It can warn of wear and growing errors so you can replace a drive in time, but it cannot predict sudden controller or power failures. Backups prevent data loss.

Is RAID the same as a backup? No. RAID protects uptime when a drive fails, but deletions, corruption and ransomware are written to all drives at once. Keep separate, isolated backups.

What is the 3-2-1 backup rule? Keep three copies of your data, on two different media types, with one copy offsite. Guidance citing NIST also recommends isolating backups and testing restores regularly.