On 09/09/2023 12:26, David T-G wrote:
Hi, all --
...and then David T-G home said...
% Hi, all --
%
% After a surprise reboot the other day, I came home to find diskfarm's
% RAID5 arrays all offline with all disks marked as spares. wtf?!?
[snip]
Wow ... I'm used to responses pointing out either what I've left
out or how stupid my setup is, but total silence ... How did I
offend and how can I fix it?
Sorry, it's usually me that's the quick response, everyone else takes
ages, and I'm feeling a bit burnt out with life in general at the moment.
I sure could use advice on the current hangup before perhaps just
destroying my entire array with the wrong commands ...
I wonder if a controlled reboot would fix it. Or just do a --stop
followed by an assemble. The big worry is the wildly varying event
counts. Do your arrays have journals.
With fingers crossed,
:-D
If the worst comes to the worst, try a forced assemble with the minimum
possible drives (no redundancy). Pick the drives with the highest event
counts. You can then re-add the remaining ones if that works.
Iirc this is actually not uncommon and it shouldn't be hard to recover
from. I really ought to go through the archives, find a bunch of
occasions, and write it up.
The only real worry is that the varying event counts mean that some data
corruption is likely. Recent files, hopefully nothing important. One
thing that's just struck me, this is often caused by a drive failing
some while back, and then a glitch on a second drive brings the whole
thing down. When did you last check your array was fully functional?
Cheers,
Wol
