ZFS question

[Quartz]

> I also confirmed that **reads** from the now-busted pool do block/wait
> indefinitely

Yeah, and that's the kicker. When my pool dies, I can't even check its 
status to figure out what the hell went wrong.


>(for some reason people
> think kill -9 will always cause a kernel to unlock/relinquish a thread,
> which is not the case)).

Oh believe me, from years of experience with linux/mac I wish it were true.


>For whatever reason "ls -l /pool" did return
> results, but I imagine these may be being returned from some caching
> mechanism within the kernel (either VFS or ZFS ARC, not sure).

I've noticed this too a couple times, but it only appears to happen for 
the top directory or so. Copy a folder with a lot of nested subfolders 
and then test with 'ls -R', that will kill it guaranteed.

______________________________________
it has a certain smooth-brained appeal