I have two 2TB Samsung 870 EVO SATA SSDs. After an unexpected reboot, Event Viewer began showing thousands of errors such as "The device, \Device\Harddisk2\DR2, has a bad block." CHKDSK also reports bad clusters and seems to find more whenever I run it again. CrystalDiskInfo reports both SSDs as "Good" with 99% health, and Samsung Magician also shows no obvious problem. The drives have about 32TB and 19TB written, and they are different ages. My other NVMe drives appear normal. I've also had crashes including CLOCK_WATCHDOG_TIMEOUT and VIDEO_SCHEDULER_INTERNAL_ERROR. Both SATA SSDs may be connected to the same PSU power cable, so I'm wondering whether the drives are actually failing or whether the shared power connection, SATA cables, memory, motherboard, or another component is causing the errors.
3 Answers
The two crash codes do not specifically prove that the SSDs are responsible. CLOCK_WATCHDOG_TIMEOUT usually points to a CPU, memory, power, firmware, or motherboard stability problem, while VIDEO_SCHEDULER_INTERNAL_ERROR can involve the GPU or its driver. Check firmware and drivers, and inspect the Windows crash dumps if available. If errors continue after changing cables and testing memory, test each SSD separately in another system or with another SATA controller; replace any drive that continues producing bad-block errors.
Since both affected drives are SATA, check the shared parts first. Replace both SATA data cables, try different motherboard ports, and if possible use a different PSU power lead rather than the same daisy-chained connector. Also test the RAM with a bootable memory tester for many hours. Faulty memory, cabling, power, or a motherboard controller can produce filesystem corruption that looks like failing drives.
The 99% figure mainly reflects estimated wear and write endurance; it does not guarantee that every read, controller, cable, or memory block is working correctly. SMART health thresholds are manufacturer-defined and can remain green even when a drive is malfunctioning. Repeated CHKDSK repairs are a reason to treat the data as at risk, regardless of the percentage. Back up anything important before doing more testing.
Related Questions
Lenovo Thinkpad Stuck In Update Loop Install FilterDriverU2_Reload