[Klug-general] What you dont do is...

[Karl Lattimer]

Been there myself Kevin, simulated a failure and took out the wrong disk
when the server was running, then in my panic took out another disk that
i believed to be the correct one, then oh bugger that was the wrong one
too, tried to get the raid to re-admit the disks into the array with no
luck and then I managed to stupidly remove more disks without thinking
about any consequences, panic destroys raid, you must be at one with the
raid.

Well after the raid had well and truly been screwed by panic i tried to
rebuild the server, it booted straight away with the same disks put back
in...

Then the input output errors started...

A screens worth makes you think you need a reinstall, as it turned out 5
of 8 disks were damaged physically (sata libata ain't as rugged as
scsi), we sent them back to Maxtor and all five have now returned new.
The server is now up and running with one spare hole and one spare disk,
(we only received 4 before we decided to rebuild) now i have to raid
reconf at some point in the future but i'm weighing up the pro's and
con's of having a spare drive bay and a spare drive when it comes to a
disk failing in real life.

The server has the job of providing currently almost a hundred thousand
PDF files via a web interface with the capability of storing upto 7.5
million, the act of filing cabinet compression to a 3u rack device.

I just hope the bugger continues to run and no kids manage to gain entry
to that room (as this is a school) and start unplugging disks willy
nilly.

Be grateful you haven't lost your data.

Regards, and sympathies 
Karl,