Drive errors in raidz array

[Toby Burress]

I have a system with 24 drives in raidz2.  When testing with bonnie++
it seemed to work fine (although I had to raise the arc_max to
prevent kernel panics).  However, now we're copying data to it and
dmesg is showing many errors like:

mpt0: mpt_cam_event: 0x16
mpt0: request 0xffffff80005f3840:63495 timed out for ccb 0xffffff000988f800 (req->ccb 0xffffff000988f800)
mpt0: request 0xffffff80005f1f80:63496 timed out for ccb 0xffffff00098d0800 (req->ccb 0xffffff00098d0800)
mpt0: attempting to abort req 0xffffff80005f3840:63495 function 0
mpt0: request 0xffffff8000601ee0:63497 timed out for ccb 0xffffff011edaa800 (req->ccb 0xffffff011edaa800)
mpt0: request 0xffffff80005f4ec0:63498 timed out for ccb 0xffffff011eda5800 (req->ccb 0xffffff011eda5800)
mpt0: mpt_wait_req(1) timed out
mpt0: mpt_recover_commands: abort timed-out. Resetting controller
mpt0: mpt_cam_event: 0x0
mpt0: completing timedout/aborted req 0xffffff80005f3840:63495
mpt0: completing timedout/aborted req 0xffffff80005f1f80:63496
mpt0: completing timedout/aborted req 0xffffff8000601ee0:63497
mpt0: completing timedout/aborted req 0xffffff80005f4ec0:63498

followed by

(da0:mpt0:0:1:0): READ(10). CDB: 28 0 1 23 81 6f 0 0 2b 0 
(da0:mpt0:0:1:0): CAM Status: SCSI Status Error
(da0:mpt0:0:1:0): SCSI Status: Check Condition
(da0:mpt0:0:1:0): UNIT ATTENTION asc:29,0
(da0:mpt0:0:1:0): Power on, reset, or bus device reset occurred
(da0:mpt0:0:1:0): Retrying Command (per Sense Data)

for every drive in the array.  Additionally, zpool scrub says:

 pool: backups
 state: ONLINE
status: One or more devices has experienced an unrecoverable error.  An
        attempt was made to correct the error.  Applications are unaffected.
action: Determine if the device needs to be replaced, and clear the errors
        using 'zpool clear' or replace the device with 'zpool replace'.
   see: http://www.sun.com/msg/ZFS-8000-9P
 scrub: resilver completed after 0h0m with 0 errors on Thu Jan 21 23:15:36 2010

I'm using 8.0-RELEASE-p2 on amd64.  One other thing that changed
between testing with bonnie++ and now is that I used glabel to label
the drives before I put them in the raidz array.

There is no raid controller.

Is this something anyone has seen before?  Googling around shows
some similar errors but no solutions.