Jon Green wrote:
snip
Certainly not a case of a fatal failure within minutes or hours,
although Roland's right that it can sometimes go catastrophic a lot more
quickly.

Jon
(* Sorry to bang on about it -- it's the best example I have of a fully
instrumented drive failure; usually we just swap regardless, ASAP.)

You would *not* have wanted to be in charge of the radio station on a
cruise ship where one afternoon the ships comms started playing up, with
no new drives onboard, and I spent the evening starting to get a copy of
the data off onto another machine, begged the C/eng to get me a new
drive pronto (we were alongside) which he did thank heavens, got some
sleep, looked at it in the morning only to find it was worse, and got
enough off it restart the system when the new drive came onboard. My
experience is the quality of the equipment onboard is in inverse
proportion to the amount of money the pax are paying.

That's why I said when a drive starts to go, it can go in hours. Which
is not funny when it's not part of a RAID setup.

PB