A primary database server ceased operating during normal business hours. It answered no network commands, and when I examined the physical machine in the datacenter, its power indicators were dark; the remote management interface likewise gave no reply.

The database is important for the company, so the repair could not wait. I prepared a replacement motherboard from our stock and gathered the necessary tools and an electrostatic discharge (ESD) wrist strap.

I disconnected all power cables from the failed server, drew it from the rack, and opened the chassis. One by one I disconnected the internal cables, removed the central processing units (CPUs), memory modules, and expansion cards, and lifted out the failed motherboard. The new board I seated in its place, reinstalled the CPUs, memory modules, and expansion cards, connected all internal cables, closed the chassis, and returned the server to the rack with its power and network cables.

I started the server. It completed the power-on self-test (POST) without errors; the operating system loaded correctly; the database started normally. I updated the asset management system with the new hardware details. The service was fully restored in two hours.