zfs discuss - Dec 2010 - copy complete zpool via zfs send/recv

Hi,

I want to move all the ZFS fs from one pool to another, but I don''t
want
to "gain" an extra level in the folder structure on the target pool.
On the source zpool I used zfs snapshot -r tank at moveTank on the root fs 
and I got a new snapshot in all sub fs, as expected.

Now, I want to use zfs send -R tank at moveTank | zfs recv targetTank/... 
which would place all zfs fs one level down below targetTank.
Overwriting targetTank is not an option, since the zfs fs musn''t exist 
prior to zfs recv.

Now, I could run a little script that runs through the zpool and 
transfers each zfs fs over individually, but I hoped that there would be 
another way.

Cheers,
budy

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20101217/c4ebed09/attachment.html>

> From: zfs-discuss-bounces at opensolaris.org [mailto:zfs-discuss-
> bounces at opensolaris.org] On Behalf Of Stephan Budach
> 
> Now, I want to use zfs send -R tank at moveTank | zfs recv targetTank/...
> which would place all zfs fs one level down below targetTank.
> Overwriting targetTank is not an option, since the zfs fs musn''t
exist
prior to
> zfs recv.
You lost me in that one.

If there is a zfs filesystem at the recipient...  and you don''t want to
receive into a subdirectory of it ... then you have to overwrite it.  And I
don''t see any reason not to overwrite it.

You say the destination filesystem mustn''t exist prior to receive... 
But
did you know of the -F option?  It forces the overwrite, if you want to
receive and overwrite an existing filesystem.

What am I missing?  Or did I just answer it?

Am 18.12.10 05:44, schrieb Edward Ned Harvey:
>> From: zfs-discuss-bounces at opensolaris.org [mailto:zfs-discuss-
>> bounces at opensolaris.org] On Behalf Of Stephan Budach
>>
>> Now, I want to use zfs send -R tank at moveTank | zfs recv
targetTank/...
>> which would place all zfs fs one level down below targetTank.
>> Overwriting targetTank is not an option, since the zfs fs
musn''t exist
> prior to
>> zfs recv.
> You lost me in that one.
>
> If there is a zfs filesystem at the recipient...  and you don''t
want to
> receive into a subdirectory of it ... then you have to overwrite it.  And I
> don''t see any reason not to overwrite it.
>
> You say the destination filesystem mustn''t exist prior to
receive...  But
> did you know of the -F option?  It forces the overwrite, if you want to
> receive and overwrite an existing filesystem.
>
> What am I missing?  Or did I just answer it?
>
Ehh... well... you answered it... sort of. ;)
I think I simply didn''t dare to overwrite the root zfs on the 
destination zpool with -F, but of course you''re right, that this is the
way to go.

Thanks,
budy
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20101218/d9b710b2/attachment.html>

> From: Stephan Budach [mailto:stephan.budach at jvm.de]
> 
> Ehh. well. you answered it. sort of. ;)
> I think I simply didn''t dare to overwrite the root zfs on the
destination
zpool
> with -F, but of course you''re right, that this is the way to go.
What are you calling the root zfs on the destination?
You''re not trying to overwrite / are you?  That would ... admittedly
... not
be so straightforward.  But I don''t think it''s impossible.

Am 18.12.10 15:14, schrieb Edward Ned Harvey:
>> From: Stephan Budach [mailto:stephan.budach at jvm.de]
>>
>> Ehh. well. you answered it. sort of. ;)
>> I think I simply didn''t dare to overwrite the root zfs on the
destination
> zpool
>> with -F, but of course you''re right, that this is the way to
go.
> What are you calling the root zfs on the destination?
> You''re not trying to overwrite / are you?  That would ...
admittedly ... not
> be so straightforward.  But I don''t think it''s
impossible.
>
The root zfs, to me, is the fs that gets created once you create the 
zpool. So, if I create the zpool tank, I also get the zfs fs tank, no?

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20101218/fc79c9b3/attachment.html>

2010/12/18 Stephan Budach <stephan.budach at jvm.de>
>  Am 18.12.10 15:14, schrieb Edward Ned Harvey:
>
>  From: Stephan Budach [mailto:stephan.budach at jvm.de <stephan.budach
at jvm.de>]
>
> Ehh. well. you answered it. sort of. ;)
> I think I simply didn''t dare to overwrite the root zfs on the
destination
>
>  zpool
>
>  with -F, but of course you''re right, that this is the way to go.
>
>  What are you calling the root zfs on the destination?
> You''re not trying to overwrite / are you?  That would ...
admittedly ... not
> be so straightforward.  But I don''t think it''s
impossible.
>
>
>  The root zfs, to me, is the fs that gets created once you create the
> zpool. So, if I create the zpool tank, I also get the zfs fs tank, no?
>
> Yes, but zfs receive can put received data only to another pool. You cannot
zfs receive to RAW  disk
>
> _______________________________________________
> zfs-discuss mailing list
> zfs-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20101218/4e7d4e24/attachment-0001.html>

Am 18.12.10 15:54, schrieb eXeC001er:
>
>
> 2010/12/18 Stephan Budach <stephan.budach at jvm.de 
> <mailto:stephan.budach at jvm.de>>
>
>     Am 18.12.10 15:14, schrieb Edward Ned Harvey:
>>>     From: Stephan Budach [mailto:stephan.budach at jvm.de]
>>>
>>>     Ehh. well. you answered it. sort of. ;)
>>>     I think I simply didn''t dare to overwrite the root zfs
on the destination
>>     zpool
>>>     with -F, but of course you''re right, that this is the
way to go.
>>     What are you calling the root zfs on the destination?
>>     You''re not trying to overwrite / are you?  That would ...
admittedly ... not
>>     be so straightforward.  But I don''t think it''s
impossible.
>>
>     The root zfs, to me, is the fs that gets created once you create
>     the zpool. So, if I create the zpool tank, I also get the zfs fs
>     tank, no?
>
> Yes, but zfs receive can put received data only to another pool. You 
> cannot zfs receive to RAW  disk
>
>
Well, in the end it is of course as Edward pointed out. Imagine a zpool 
called tank with 4 ZFS volumes on it:

tank
tank/fs1
tank/fs2
tank/fs3

This sums up to 4 volumes (maybe I should have used volumes, rather then 
fs).

I wanted to transfer the complete zpool tank onto another zpool called 
backupTank by using zfs send -R tank at snap | zfs recv backupTank, which 
raised the error message that I can''t overwrite the existing volume. 
This can of course be overwritten by using -F, such as that it would 
replace everything on backupTank with the contents of tank - this is 
what I didn''t dare, since for some reason, I thought that the volume 
tankBackup would get destroyed, but in the end it''s not the volume that
gets destroyed, but it''s contents gets overwritten.

Cheers,
budy
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20101218/9da9ae31/attachment.html>