Re: [zfs-discuss] Should I report this as a bug?

2009-02-07 Thread Kees Nuyt
On Wed, 04 Feb 2009 11:12:55 -0500, Kyle McDonald
kmcdon...@egenera.com wrote:

I jumpstarted my machine with sNV b106, and installed with ZFS root/boot.
It left me at a shell prompt in the JumpStart environment, with my ZFS 
root on /a.

I wanted to try out some things that I planned on scripting for the 
JumpStart to run, one of these waas creating a new ZFS pool from the 
remaining disks. I looked at the zpool create manpage, and saw this it 
had a -R altroot option, and the exact same thing had just worked for 
me with 'dladm aggr-create' so I thought I'd give that a try.

If the machine had been booted normally, my ZFS root would have been /, 
and a 'zpool create zdata0 ...' would have defaulted to mounting the new 
pool as /zdata0 right next to my ZFS root pool /zroot0. So I expected 
'zpool create -R /a zdata0 ...' to set the default mountpoint for the 
pool to /zdata0 with a temporary altroot=/a.

I gave it a try, and while it created the pool it failed to mount it at 
all. It reported that /a wasn't empty.

'zpool list', and 'zpool get all' show the altroot=/a. But 'zfs  get 
all  zdata0' shows the mountpoint=/a also, not the default of /zdata0.

Am I expecting the wrong thing here? or is this a bug?

My guess is /a is occupied by the mount of the just
installed root pool.

You'll have to create a new mountpoint, something like /b,
and have your zdata0 pool mount there temporarily.

 -Kyle
-- 
  (  Kees Nuyt
  )
c[_]
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


[zfs-discuss] Should I report this as a bug?

2009-02-04 Thread Kyle McDonald
I jumpstarted my machine with sNV b106, and installed with ZFS root/boot.
It left me at a shell prompt in the JumpStart environment, with my ZFS 
root on /a.

I wanted to try out some things that I planned on scripting for the 
JumpStart to run, one of these waas creating a new ZFS pool from the 
remaining disks. I looked at the zpool create manpage, and saw this it 
had a -R altroot option, and the exact same thing had just worked for 
me with 'dladm aggr-create' so I thought I'd give that a try.

If the machine had been booted normally, my ZFS root would have been /, 
and a 'zpool create zdata0 ...' would have defaulted to mounting the new 
pool as /zdata0 right next to my ZFS root pool /zroot0. So I expected 
'zpool create -R /a zdata0 ...' to set the default mountpoint for the 
pool to /zdata0 with a temporary altroot=/a.

I gave it a try, and while it created the pool it failed to mount it at 
all. It reported that /a wasn't empty.

'zpool list', and 'zpool get all' show the altroot=/a. But 'zfs  get 
all  zdata0' shows the mountpoint=/a also, not the default of /zdata0.

Am I expecting the wrong thing here? or is this a bug?

 -Kyle

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss