An apparatus, program product and method of replacing a failed hardware device in a computer each rely solely on program code and/or other computer facilities that are ensured of being available in the computer during a concurrent maintenance operation, so that, even in the event that a failure occurs in such a hardware device, successful performance of the concurrent maintenance operation is ensured. For example, the initiation of power up and power down functions necessary to permit replacement of a failed device may be performed through a control panel or other similar facility in a computer that is continuously available when a computer is in a fully or partially active and powered-on state. Also, a concurrent replacement operation may be supported that automatically detects a lock-up condition resulting from a failed attempt to access data using a failed hardware device, and that, upon replacement of the device with a suitable replacement device, automatically associates a resource that was previously associated with the failed device with the replacement device such that the failed attempt to access data may be automatically resumed, thereby automating the recovery from the lock-up condition.