zfs discuss - May 2008 - [storage-discuss] ZFS and fibre channel issues

The ZFS crew might be better to answer this question. (CC''d here)

       --jc

William Yang wrote:
> I am having issues creating a zpool using entire disks with a fibre 
> channel array.  The array is a Dell PowerVault 660F.
> When I run "zpool create bottlecap c6t21800080E512C872d14 
> c6t21800080E512C872d15", I get the following error:
> invalid vdev specification
> use ''-f'' to override the following errors:
> /dev/dsk/c6t21800080E512C872d0s0 is part of active ZFS pool nalgene. 
> Please see zpool(1M).
> I was able to create pool nalgene by using entire disks, but this was 
> awhile back.  I have a feeling one of the patches I applied after 
> creating nalgene broke something.  I am currently using Solaris 10 
> SPARC 8/07 kernel 127111-11.
>  
> Also, if I append s0 to the disk name (i.e. c6t21800080E512C872d14s0), 
> then I can create the new zpool.  Any ideas?
>  
> William Yang
> ------------------------------------------------------------------------
>
> _______________________________________________
> storage-discuss mailing list
> storage-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/storage-discuss
>

Hi, William,

  You didn''t mention c6t21800080E512C872d0s0 in your command line,

maybe typo of c6t21800080E512C872d14s0?

  # zpool create bottlecap c6t21800080E512C872d14
c6t21800080E512C872d15 

  The warrning looks like by the remain label info may reside on you
disk.

Could you see output by ''zdb -l
/dev/dsk/c6t21800080E512C872d14s0'' ? it
should

has something related to nalgene. But anyway, if you have re-used that
disk to create

new pool, I suspect this issue has gone.

  - Regards,

Jeff Cheeney wrote:

The ZFS crew might be better to answer this question. (CC''d here)

       --jc

William Yang wrote:

I am having issues creating a zpool using entire disks with a fibre 
channel array.  The array is a Dell PowerVault 660F.
When I run "zpool create bottlecap c6t21800080E512C872d14 
c6t21800080E512C872d15", I get the following error:
invalid vdev specification
use ''-f'' to override the following errors:
/dev/dsk/c6t21800080E512C872d0s0 is part of active ZFS pool nalgene. 
Please see zpool(1M).
I was able to create pool nalgene by using entire disks, but this was 
awhile back.  I have a feeling one of the patches I applied after 
creating nalgene broke something.  I am currently using Solaris 10 
SPARC 8/07 kernel 127111-11.

Also, if I append s0 to the disk name (i.e. c6t21800080E512C872d14s0), 
then I can create the new zpool.  Any ideas?

William Yang
------------------------------------------------------------------------

_______________________________________________
storage-discuss mailing list
storage-discuss-xZgeD5Kw2fzokhkdeNNY6A@public.gmane.org
http://mail.opensolaris.org/mailman/listinfo/storage-discuss

_______________________________________________
zfs-discuss mailing list
zfs-discuss-xZgeD5Kw2fzokhkdeNNY6A@public.gmane.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss

-- 
Regards,

Robin Guo, Xue-Bin Guo
Solaris Kernel and Data Service QE,
Sun China Engineering and Reserch Institute
Phone: +86 10 82618200 +82296
Email: robin.guo-xsfywfwIY+M@public.gmane.org
Blog: http://blogs.sun.com/robinguo


_______________________________________________
storage-discuss mailing list
storage-discuss-xZgeD5Kw2fzokhkdeNNY6A@public.gmane.org
http://mail.opensolaris.org/mailman/listinfo/storage-discuss

Hi William,
comments below (after snipping)


William Yang wrote:
> No typo, which is why I think this is a bug somewhere.  d14 has never 
> been a member of nalgene.  d0 is an active disk in pool nalgene.
>  
> My best guess is that there is an incorrect assumption somewhere that 
> assumes the disk is always d0 (probably because it is more common to 
> have single-disk targets rather than multidisk targets?).  Even running 
> format gives a warning:
> hostname:~# format /dev/rdsk/c6t21800080E512C872d14s0
> selecting /dev/rdsk/c6t21800080E512C872d14s0
> [disk formatted]
> /dev/dsk/c6t21800080E512C872d0s0 is part of active ZFS pool nalgene. 
> Please see zpool(1M).
>  
> (The same warning occurs if I use s2 instead of s0.)
Yes, that does look like a bug, please file one on bugs.opensolaris.org

....
>>     William Yang wrote:
>>       
>>>     I am having issues creating a zpool using entire disks with a
fibre
>>>     channel array.  The array is a Dell PowerVault 660F.
>>>     When I run "zpool create bottlecap c6t21800080E512C872d14 
>>>     c6t21800080E512C872d15", I get the following error:
>>>     invalid vdev specification
>>>     use ''-f'' to override the following errors:
>>>     /dev/dsk/c6t21800080E512C872d0s0 is part of active ZFS pool
nalgene.

What you''re seeing here is the interposition of libdiskmgmt,
which is what we use to figure out whether a device is in
use by a filesystem when you''re trying to do a potentially
data-destructive operation on it.

Based on my limited investigation so far, it appears that
one layer is making an assumption about your disk based on
the devid:

         path=''/dev/dsk/c6t21800080E512C872d14s0''
           devid=''id1,ssd at x21000080e512c872/a''


I don''t have a multi-lun per target system to check this,
so you''ll have to do this - when you log your bug could
you please include the output from

prtconf -v

and

zdb -l /dev/dsk/.....  for each of the luns attached
to your targets on this controller.



James C. McPherson
--
Senior Kernel Software Engineer, Solaris
Sun Microsystems
http://blogs.sun.com/jmcp	http://www.jmcp.homeunix.com/blog