Patient records server: SMART alert, and the disk replaced this morning is doing nothing

Polyclinique Les Oliviers (Meknes) keeps its patient records on srv-dossiers-01, a physical Rocky Linux 9 server with two disks in software RAID 1 (/dev/md0, mdadm) holding /srv/dossiers. The system itself is on a separate SSD.

This morning, hardware monitoring reported SMART errors on a disk. The maintenance technician found that a disk had been “failed” for several days, hot-swapped it at 08:30 with a new disk of the same model, then left: “The RAID will rebuild by itself.”

An hour later, nothing seems to have changed. If the remaining disk fails, the records of 40,000 patients are lost. You are the system administrator: assess the risk and restore redundancy.