Timelines · Maker advisories

Storage Failures Hall of Shame

Drives that lost data, stopped working after a set number of hours, crashed the computer, dropped off it or slowed down, each taken from the advisory the drive’s own seller published, with the fixed firmware where the seller names one. Ranked by what the seller says happens, worst first, and not by anyone’s guess at how many drives were hit.

The board SSD Endurance Leaderboard All Timelines

How to read it

8 of these advisories tie the failure to a count of power-on hours, the time a drive has spent switched on: once a drive has run that long on the old firmware, it fails. HPE’s advisory puts the soonest at 4,700 power-on hours, which is 195 days, 20 hours of running without a break. Cisco, Dell and HPE each published an advisory for drives that fail at 40,000 hours, 4 years, 206 days, 16 hours, which Cisco rounds to 4.5 years.

The board ranks by harm first, in an order we chose: loses data, drive stops working, crashes the computer, drops off the computer and slows down. A row’s harm follows the seller’s own words, so two advisories can describe the same symptom and still sit in different tiers: Cisco and Dell both say the drive reports 0 GB and goes offline, but only Cisco says data is lost. Inside a harm, a defect with an hour count ranks ahead of one without, fewest hours first, since a drive that fails sooner leaves less time to update it.

Hours until the drive fails8 advisories

Only the advisories that give an hour count. Tap a bar to jump to its row.

Power-on hours until the drive fails

Loses data
Drive stops working
Drops off the computer
Each year of running day and night

Linear scale from zero. Each bar is the hour count a seller’s advisory gives for the failure; the dotted lines mark 365-day years of running without a break. Advisories that give no hour count, or that say the drive only slows down, are on the board below, not on this chart.

The board23 rows

Ranked by harm, then by hours to failure. Tap a heading to sort by it, or filter by seller and interface.

Rank	Seller	Drives	Interface	What happens	At power-on hours	In the seller’s words, our summary	Fixed firmware	Seller source
1	HPE	Certain HPE NVMe SSD models	NVMe	Loses data	4,700 195 days, 20 hours	The drive fails at that hour count, and HPE says neither the drive nor its data can be recovered.	4ICRHPK3	HPE customer bulletin
2	HPE	Certain HPE SAS SSD models	SAS	Loses data	32,768 3 years, 270 days, 8 hours	The drive fails at that hour count, and HPE says the data on it is lost.	HPD8	HPE SAS SSD remediation guide
3	Cisco	Certain Cisco UCS SAS SSD models	SAS	Loses data	40,000 4 years, 206 days, 16 hours	The drive reports 0 GB of storage, goes offline and becomes unusable, and Cisco says data is lost.	C405	Cisco field notice FN70545
3	HPE	Certain HPE SAS SSD models	SAS	Loses data	40,000 4 years, 206 days, 16 hours	The drive fails at that hour count, and HPE says the data on it is lost. HPE says the defect is not unique to HPE.	HPD7	HPE SAS SSD remediation guide
5	HPE	Certain HPE SAS hard drives, 4 TB to 14 TB	SAS	Loses data	65,536 7 years, 175 days, 16 hours	The hard drive fails at that hour count and cannot be recovered, and HPE warns of data loss.	HPD3, HPD4, HPD5, HPD7	HPE customer bulletin
5	HPE	Certain HPE SATA hard drives, 4 TB to 14 TB	SATA	Loses data	65,536 7 years, 175 days, 16 hours	The hard drive fails at that hour count and cannot be recovered, and HPE warns of data loss.	HPG4, HPG5, HPG6, HPG8	HPE customer bulletin
7	Dell	Dell Intel S4510 and S4610 SSDs, 1.92 TB and 3.84 TB	SATA	Loses data	not stated	The drive can stop answering the host or fail early, and after a power cycle Dell says its data cannot be recovered.	DL63, XCV10110	Dell knowledge base article
8	Lenovo	Kioxia BG5 SSDs in ThinkPad, ThinkCentre and ideacentre PCs	NVMe	Loses data	not stated	Large installs or file copies can hang the PC, then the drive is never detected again and its data is lost.	1107ANLA	Lenovo support article
9	Dell	Dell Enterprise SAS SSDs LT0200MO to LT0800WM, 9 models	SAS	Drive stops working	40,000 4 years, 206 days, 16 hours	The drive reports 0 GB of storage, goes offline and cannot be used after a power cycle.	D417	Dell knowledge base article
10	Dell	Dell Toshiba PM2, PM2+ and PM2R SAS SSDs, 9 part numbers	SAS	Drive stops working	70,000 7 years, 361 days, 16 hours	The drive locks after a power cycle, takes no reads or writes and shows as 0 GB, and a firmware update unlocks it.	A3B4, A4B4, A5B4	Dell knowledge base article
11	Dell	Dell Enterprise Intel SATA SSDs, 20 part numbers from 240 GB to 7.6 TB	SATA	Drive stops working	not stated	The drive may fail early, and the server may show it as failed or missing.	DL7A	Dell knowledge base article
12	Dell	Micron 7450 and 7500 NVMe SSDs sold by Dell	NVMe	Drive stops working	not stated	The drive may fail to initialize, and Dell says a drive that hits this cannot be recovered.	E2MU300, E3MQ005	Dell knowledge base article
13	Dell	Samsung PM883a SATA SSDs sold by Dell	SATA	Drive stops working	not stated	In rare cases the drive reports ERRORMOD and 0 GB, and the failed drive may become inaccessible.	HG58	Dell knowledge base article
14	Lenovo	ThinkSystem Samsung PM1653 and PM1655 SAS SSDs on firmware CA00	SAS	Drive stops working	not stated	Garbage collection falls behind, the drive times out, and a RAID controller marks it failed for good.	CA01	Lenovo technical tip
15	HPE	HPE NVMe SSDs on firmware MPK77H5Q, MPK7725Q or HPK5	NVMe	Crashes the computer	not stated	The drive may cause an uncorrectable machine check exception on HPE Gen9 to Gen11 servers.	MPK7825Q, MPK78H5Q, HPK6	HPE customer advisory
16	Sandisk	2 TB WD_BLACK SN770 and SN770M, WD Blue SN580 and SN5000, SANDISK Extreme M.2	NVMe	Crashes the computer	not stated	A Host Memory Buffer issue may crash Windows 11 24H2 with a blue screen, and Microsoft may hold back the 24H2 upgrade until the drive is updated.	731130WD, 281050WD, 291020WD	Sandisk support article
17	Cisco	Micron M500IT SATA SSD in Cisco UCS 6324 and 6332 fabric interconnects	SATA	Drops off the computer	28,224 3 years, 81 days	The drive becomes unresponsive until a power cycle, and Cisco says no data is lost.	CZ03, MC03, MU05	Cisco field notice FN72028
17	Cisco	Micron M500IT SATA SSD in Cisco Nexus 9000 and 3000 switches	SATA	Drops off the computer	28,224 3 years, 81 days	The drive stops responding and the switch bootflash goes read-only, and it comes back after a power cycle for about six more weeks.	MC03, MU05	Cisco field notice FN72150
19	HPE	Certain HPE NVMe SSD models	NVMe	Drops off the computer	not stated	After resets the drive may show as 0 GB and stay undetected until the host is rebooted or power cycled.	EPK76H3Q, MPK76H5Q, MPK7625Q, HPK6	HPE customer advisory
20	Lenovo	ThinkSystem M.2 Micron 7450 PRO 480 GB and 960 GB on firmware E2MU300	NVMe	Drops off the computer	not stated	After under 1% of power cycles the server logs the drive as faulted and its RAID volume can go offline until the next power cycle.	not named	Lenovo technical tip
21	Lenovo	ThinkSystem Solidigm S4520 7.68 TB SATA SSDs	SATA	Drops off the computer	not stated	After a warm reboot from ESXi the drive may not be seen by the server controller, UEFI or ESXi.	not named	Lenovo technical tip
22	Sandisk	SANDISK Extreme and Extreme PRO Portable SSD, WD My Passport SSD, made before 2023	USB	Drops off the computer	not stated	Some drives unexpectedly disconnect from the computer. Sandisk says to back up and verify data before updating.	not named	Sandisk support article
23	Dell	Micron 5100 and 5200 SATA SSDs sold by Dell	SATA	Slows down	65,535 7 years, 175 days, 15 hours	A 16-bit log counter tops out and the drive floods its own debug log, so reads and writes slow; Dell says no data is lost.	L013, F006	Dell knowledge base article

What it will not tell you

It does not tell you how many drives failed. No advisory here gives a count, so the board ranks by what the seller says happens, which is a ranking of harm and not of how widespread a defect was. Cisco, Dell and HPE each give 40,000 hours, and Cisco and Dell describe the same symptom. HPE says its defect came from an SSD supplier and is not unique to HPE, but none of these pages says whether the three are one part.

A seller on this board is the one that published the advisory, which is not always the company that made the drive or wrote its firmware. The board holds only defects whose seller still publishes its own advisory, so failures known only from forums, news stories or a seller page that is gone are left out, however well known they are. The update protects a drive that has not reached the defect yet; Cisco says a drive that already reached 40,000 hours is unusable and must be replaced. Each row links the seller’s page it was read from on November 3, 2026.

The guides behind this data

Does an SSD Fail With No Warning?
Data-center studies show SSD errors repeat and the drive counts them, but the counters sit in a log few people read and the usual reliability number misses it.
Flash Foundations, Part 5: The Controller
Meet the processor coordinating the host, NAND, and the firmware inside an SSD.
Do SSDs Wear Out in a Couple of Years?
My SMART data from a system drive, a test bench, new and used drives, against the ratings and field studies: a desktop SSD wears far slower than the myth says.