zfs discuss - Dec 2006 - Can''t destroy corrupted pool

Ok, so I''m planning on wiping my test pool that seems to have problems 
with non-spare disks being marked as spares, but I can''t destroy it:

# zpool destroy -f zmir
cannot iterate filesystems: I/O error

Anyone know how I can nuke this for good?

Jim
 
 
This message posted from opensolaris.org

BTW, I''m also unable to export the pool -- same error.

Jim
 
 
This message posted from opensolaris.org

Here''s the truss output:

402:    ioctl(3, ZFS_IOC_POOL_LOG_HISTORY, 0x080427B8)  = 0
402:    ioctl(3, ZFS_IOC_OBJSET_STATS, 0x0804192C)      = 0
402:    ioctl(3, ZFS_IOC_DATASET_LIST_NEXT, 0x0804243C) = 0
402:    ioctl(3, ZFS_IOC_DATASET_LIST_NEXT, 0x0804243C) Err#3 ESRCH
402:    ioctl(3, ZFS_IOC_SNAPSHOT_LIST_NEXT, 0x0804243C) Err#5 EIO
402:    fstat64(2, 0x08041400)                  = 0
cannot iterate filesystems402:  write(2, " c a n n o t   i t e r a"..,
26)      = 26
: 402:  write(2, " :  ", 2)                     = 2
I/O error402:   write(2, " I / O   e r r o r", 9)           = 9

I did take one snapshot before my last disk spindown. Should I try nuking it?

Jim
 
 
This message posted from opensolaris.org

Nevermind:

# zfs destroy zmir at 2006-11-30-10:28
cannot open ''zmir at 2006-11-30-10:28'': I/O error

Jim
 
 
This message posted from opensolaris.org

You are likely hitting:

6397052 unmounting datasets should process /etc/mnttab instead of traverse DSL

Which was fixed in build 46 of Nevada.  In the meantime, you can remove
/etc/zfs/zpool.cache manually and reboot, which will remove all your
pools (which you can then re-import on an individual basis).

- Eric

On Mon, Dec 11, 2006 at 06:58:22AM -0800, Jim Hranicky
wrote:
> Ok, so I''m planning on wiping my test pool that seems to have
problems
> with non-spare disks being marked as spares, but I can''t destroy
it:
> 
> # zpool destroy -f zmir
> cannot iterate filesystems: I/O error
> 
> Anyone know how I can nuke this for good?
> 
> Jim
>  
>  
> This message posted from opensolaris.org
> _______________________________________________
> zfs-discuss mailing list
> zfs-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
--
Eric Schrock, Solaris Kernel Development       http://blogs.sun.com/eschrock

> You are likely hitting:
> 
> 6397052 unmounting datasets should process
> /etc/mnttab instead of traverse DSL
> 
> Which was fixed in build 46 of Nevada.  In the
> meantime, you can remove
> /etc/zfs/zpool.cache manually and reboot, which will
> remove all your
> pools (which you can then re-import on an individual
> basis).
I''m running b51, but I''ll try deleting the cache.

Jim
 
 
This message posted from opensolaris.org

This worked. 

I''ve restarted my testing but I''ve been fdisking each drive
before I
add it to the pool, and so far the system is behaving as expected
when I spin a drive down, i.e., the hot spare gets automatically used. 
This makes me wonder if it''s possible to ensure that the forced
addition of a drive to a pool wipes the pool of any previous data, 
especially any zfs metadata.

I''ll keep the list posted as I continue my tests.

Jim
 
 
This message posted from opensolaris.org