I just ran zpool scrub on an active pool on an x4170 running S10U7 with the
latest patches and iostat immediately dropped to 0 for all the pool devices and
all processes associated with that device where hard locked, e.g., kill -9 on a
zpool status processes was ineffective. However, other zpool on the system, such
as the root pool, continued to work.
Neither init 6 nor reboot where able to take the system all the way down, though
reboot did get further. After a hard reset the system came backup cleanly and a
subsequent zpool scrub succeeded, but I am now concerned about when it is safe
to run a scrub. The most interesting pool action taken before the scrub was
attempted was to append another mirror vdev to the pool which already had 2
mirrors in it.
Are there known bugs I can track with regards to zpool scrub locking up a
system, and if it happens again is there any useful information I can gather
before or during the system reset to help track this down further?
Thanks.
--
This message posted from opensolaris.org
