On 09/ 2/11 11:58 PM, Edward Ned Harvey wrote:
The offline command never completes and all I/O activity on the pool is
suspended. The command could not be killed and a power cycle was required..
From: zfs-discuss-boun...@opensolaris.org [mailto:zfs-discuss-
boun...@opensolaris.org] On Behalf Of Ian Collins
I have an S11 Express system with a pool with an SSD split between ZIL
and cache. My friendly power company donated a free spike that killed
the SSD and one of the pool drives.
I was able to successfully zpool remove the cache slice from the pool,
but any attempt at removing the log slice locks up the system.
Any ideas? I can run with the pool in a degraded state, but how will I
replace the log if it can't be removed?
IIRC, even though the device is faulted, you have to offline it before you
can remove it. I think. (As in, zpool offline, followed by zpool remove).
Still... It shouldn't lock up the system. I assume you're doing "zpool
remove" log device, right? And when you do that, it locks up the system?
zfs-discuss mailing list