Every case is different, but the failures repeat. These are the shapes they come in: what arrived, what had already been tried, what was done about it and what came back.
Array ran degraded for weeks unnoticed. A second member dropped, and the rebuild was started against the stale disk before anyone called. Members were imaged individually and the array assembled virtually from the pre-rebuild state.
Drive clicked for several days of continued attempts before it came in. Heads were replaced from a matched donor and the platters imaged head by head, with the damaged surface zones read last.
Error 528 after a power event, with logs that would not replay. The underlying volume was imaged first, the store repaired against the image, then split back out into per-mailbox PST files.
Backup job had been failing silently for months. No decryption was possible, so recovery came from partially deleted shadow copies and pre-encryption originals still sitting in unallocated space.
Connector sheared off with the board cracked through. The NAND package was intact, so the chip was read directly and the translation layer reconstructed outside the device.
Two-bay NAS stopped powering on and the vendor advised a replacement unit and a restore from backup that did not exist. The members were imaged and the volume mounted in-lab.
Describe the failure and what has already been tried. The diagnosis is free either way.
Data Recovery Lab