BrevisSalt

My friend had been running an Unraid 7 server for a long time without any major problems. The machine had an Intel motherboard, an LSI controller, and eight SATA hard drives, mostly 12TB Western Digital Ultrastar models, along with a 14TB Ultrastar parity drive. Then the parity drive failed. Replacing it should have been straightforward. He bought another 14TB Ultrastar drive that appeared almost identical to the original and installed it in the same slot. Nothing happened. There was no spin-up, no vibration, and no sign that the drive was receiving power. Thinking the replacement might simply be defective, he tried a spare 12TB drive. Exactly the same thing happened. The original parity drive, however, still spun up immediately when put back into the system. That was when he realized the problem probably wasn't two dead hard drives appearing at exactly the wrong moment. We started troubleshooting the hardware chain rather than assuming the drives were incompatible. We checked the LSI controller, SATA connections, power cables, backplane connections, drive detection, and whether the replacement drives were actually receiving power. We also compared discussions and documentation from Western Digital, Seagate, and ServeTheHome to understand differences between enterprise drives and their power requirements. One particularly important detail was the possibility of 3.3V power-disable (PWDIS) on certain enterprise SATA drives. Some drives can refuse to spin up when connected to a power source that presents 3.3V on the relevant pin, while an older drive in the same system may work normally. That can make an apparently identical replacement look completely dead even though the drive itself isn't necessarily faulty. Before buying another replacement, my friend started checking the exact drive model and power configuration rather than relying only on the capacity and product family. He also compared replacement storage hardware and server components while making sure the replacement drive matched the server's actual power and controller setup. The mystery finally made sense. The issue wasn't that Unraid suddenly disliked 14TB drives, nor that two manufacturers had coordinated the world's most inconvenient HDD prank. The replacement drives simply behaved differently with the server's existing power configuration. The experience taught my friend an important lesson about enterprise storage: two hard drives can look almost identical on the specification sheet and still behave differently in a particular server because of firmware, power features, controller compatibility, or connector configuration. When a replacement drive doesn't even spin up, the first question shouldn't always be, “Is the drive dead?” Sometimes the better question is: “Is the server actually giving this particular drive the kind of power it expects?”