and – test them on a separate system! My first major system crash was when an IBM model 95 developed a synchronous SCSI controller fault. Data was garbled writing TO disk or tape, and ungarbled reading FROM disk or tape.
All good until the controller failed and was replaced. The data on the disks were now unreadable. The data on the tapes matched that on the disks exactly – so also unreadable.
3 months work on the central file system, for 300 people, lost forever.
and – test them on a separate system! My first major system crash was when an IBM model 95 developed a synchronous SCSI controller fault. Data was garbled writing TO disk or tape, and ungarbled reading FROM disk or tape.
All good until the controller failed and was replaced. The data on the disks were now unreadable. The data on the tapes matched that on the disks exactly – so also unreadable.
3 months work on the central file system, for 300 people, lost forever.