On 23 May 2009 at 7:35, Rick Bensene wrote:
Whatever it was, it completely destroyed the drive.
The CDC drive
(can't remember the model number, unfortunately) was very well-built.
Given the time, it sounds like an 844; the only other pack drive
that I can recall from that period was the 854, which was more the
size of an IBM 2311 (not used much on 6000/Cyber 70 because it had
256 8-bit byte-sized sectors hard-wired onto it and standard sector
size for the CDC gear was 644 6-bit characters. Two 854 sectors for
one 6000 sector left much of the drive empty).
We were on a project that used farms of the 844s (over 100 was
typical) when they were still very new. (The contract had originally
been spec-ed for the new, but ill-fated 821 4-spindle fixed-disk
drive, along the lines of an 808, but the 821 performed so unreliably
and cost so much that it was scrapped. I used to have a platter from
one, but still have a head).
I used to have a memo from one of the Comsource night operators
detailing his experience with running our benchmark. It started off
something like this "I mounted all of the packs and started the
benchmark, but then got an error on unit xx. When I went to the
drive it was making a funny sound, so I moved the pack to unit yy,
which then started making a funny sound, so I got the backup pack and
mounted it on unit xx...") The eventual toll was something like 11
packs and 7 drives. The blow-by-blow account of the disaster
reminded me of the Gerard Hoffnung story about the workman filing an
accident report after he'd tried to lower a barrel of bricks from a
rooftop.
--Chuck