>From the further adventures of trying to get the broken mirror pieces glued >back together:
I tried "zfs replace data c5t3d0" to pretend I'd swapped out the "bad" disk for a new one, but zfs is too smart for me: "/dev/dsk/c5t3d0s0 is part of active ZFS pool data. Please see zpool(1M)." dd 10k of zeroes over the start of the disk, not enough, dd 1g of zeroes and I get "device is busy" when trying to do the replace command. It's offline, how can it be busy? I go take a look at the drives to see if maybe it's still flushing buffers from the dd, though I don't think that's a buffered operation. See the other drives flashing as I'm doing this while waiting for a copy to finish, but c5t3d0 is quiescent. Pull it out and put it back in to see if the system wants to see a disk change (grasping at straws, I know). Try the replace command, this time I get "cannot open 'c5t3d0': no such device in /dev/dsk", despite the fact that /var/adm/messages shows the disk info it found when it noticed the drive being plugged back in (and the "supremely detailed" device name matched what it was before). Reboot. status says "resilver completed after 0h0m with 0 errors" at the time of reboot, but c5t3d0 is still offline. replace results in "c5t3d0 is busy", ok, maybe it resilvered and just needs to be told to come back online: "warning: device 'c5t3d0' onlined, but remains in faulted state, use 'zpool replace'". status shows "degraded" and "too many errors", followed by "errors: no known data errors" and the cksum count of 0. OK, I'll try replace again, but apparently it *did* resilver, as now it thinks it's part of an active pool again. Try to force the replace with -f, but: invalid vdev specification the following errors must be manually repaired: /dev/dsk/c5t3d0s0 is part of active ZFS pool data. Please see zpool(1M). status still degraded. I'm running out of ideas. One last try: dd 600G of zeroes over the 500G drive... Surely there's a better way to reuse a drive... (though granted, in a real situation, you *should* zero the thing first...). This message posted from opensolaris.org _______________________________________________ storage-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/storage-discuss
