RAID Munging Question

[kentborg at borg.org (Kent Borg)]

On Fri, Apr 05, 2002 at 10:24:11PM -0500, christoph at linuxsoup.com wrote:
> Gordon & I simulated a disk failure on a raid1 setup at the
> last BLU meeting.  The system DID hang.  Perhaps there's a lengthy i/o
> timeout which would eventually allow the system to keep running,
> OR perhaps there are certain IDE features (DMA?) one could turn off to
> improve the recovery. But "hang" it did!

Hmmm, interesting.  

I was considering experimenting with that myself, but couldn't decide
how to do the experiment.  Just pull the power perhaps?  I figured
that would do it, but I was unsure how to put that disk back in the
array.  If one just plugs it back in, is the system smart enough to
know which disk should win and update the other one?  

How did you do it?  Makes me wish I had more hardware sitting around
so I could do more experiments without disrupting stuff I care about.


-kb, the Kent who had a busy weekend and so didn't have the chance to
do his 0->1 conversion.