zfs discuss - Dec 2009 - Doing ZFS rollback with preserving later created clones/snapshot?

Hi.

Is it possible on Solaris 10 5/09, to rollback to a ZFS snapshot,
WITHOUT destroying later created clones or snapshots?

Example:

--($ ~)-- sudo zfs snapshot rpool/ROOT at 01

--($ ~)-- sudo zfs snapshot rpool/ROOT at 02

--($ ~)-- sudo zfs clone rpool/ROOT at 02 rpool/ROOT-02

--($ ~)-- LC_ALL=C sudo zfs rollback rpool/ROOT at 01
cannot rollback to ''rpool/ROOT at 01'': more recent snapshots
exist
use ''-r'' to force deletion of the following snapshots:
rpool/ROOT at 02

So it isn''t as simple as that. But what needs to be done, to preserve
rpool/ROOT-02?

Actually, I''m not concerned (that much) with preserving the clone
rpool/ROOT-02. But I''d like to keep the contents of rpool/ROOT as
it was when I created the @02 snapshot.

Is the only possible way to create a backup of rpool/ROOT at 02 (eg.
of the snapshot directory /rpool/ROOT/.zfs/snapshots/02) and then
restore it later on (eg. backup to tape, backup to someother filesystem
using zfs send|recv, rsync, tar, ...)?

Thanks a lot,

Alexander
--
? Keine Internetzensur in Deutschland! ? http://zensursula.net ? ?
?? ? Lifestream (Twitter, Blog, ?) ??http://alexs77.soup.io/ ? ? ?
? Chat (Jabber/Google Talk) ? a.skwar at gmail.com , AIM: alexws77 ??

Yes, although it''s slightly indirect:

	- make a clone of the snapshot you want to roll back to
	- promote the clone

See ''zfs promote'' for details.

Jeff

On Fri, Dec 11, 2009 at 08:37:04AM +0100, Alexander Skwar
wrote:
> Hi.
> 
> Is it possible on Solaris 10 5/09, to rollback to a ZFS snapshot,
> WITHOUT destroying later created clones or snapshots?
> 
> Example:
> 
> --($ ~)-- sudo zfs snapshot rpool/ROOT at 01
> 
> --($ ~)-- sudo zfs snapshot rpool/ROOT at 02
> 
> --($ ~)-- sudo zfs clone rpool/ROOT at 02 rpool/ROOT-02
> 
> --($ ~)-- LC_ALL=C sudo zfs rollback rpool/ROOT at 01
> cannot rollback to ''rpool/ROOT at 01'': more recent
snapshots exist
> use ''-r'' to force deletion of the following snapshots:
> rpool/ROOT at 02
> 
> So it isn''t as simple as that. But what needs to be done, to
preserve
> rpool/ROOT-02?
> 
> Actually, I''m not concerned (that much) with preserving the clone
> rpool/ROOT-02. But I''d like to keep the contents of rpool/ROOT as
> it was when I created the @02 snapshot.
> 
> Is the only possible way to create a backup of rpool/ROOT at 02 (eg.
> of the snapshot directory /rpool/ROOT/.zfs/snapshots/02) and then
> restore it later on (eg. backup to tape, backup to someother filesystem
> using zfs send|recv, rsync, tar, ...)?
> 
> Thanks a lot,
> 
> Alexander
> --
> ? Keine Internetzensur in Deutschland! ? http://zensursula.net ? ?
> ?? ? Lifestream (Twitter, Blog, ?) ??http://alexs77.soup.io/ ? ? ?
> ? Chat (Jabber/Google Talk) ? a.skwar at gmail.com , AIM: alexws77 ??
> _______________________________________________
> zfs-discuss mailing list
> zfs-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss

Hello Jeff!

Could you (or anyone else, of course *G*) please show me how?

Situation:

There shall be 2 snapshots of a ZFS called rpool/rb-test
Let''s call those snapshots "01" and "02".

$ sudo zfs create rpool/rb-test

$ zfs list rpool/rb-test
NAME            USED  AVAIL  REFER  MOUNTPOINT
rpool/rb-test    18K   170G    18K  /rpool/rb-test

$ sudo zfs snapshot rpool/rb-test at 01
$ sudo zfs snapshot rpool/rb-test at 02
$ sudo zfs clone rpool/rb-test at 02 rpool/rb-test-02
$ sudo zfs promote rpool/rb-test-02
$ zfs list -r rpool | grep rb-test
rpool/rb-test                                              0   170G    18K
/rpool/rb-test
rpool/rb-test-02                                         18K   170G    18K
/rpool/rb-test-02
rpool/rb-test-02 at 01                                        0      -    18K
-
rpool/rb-test-02 at 02                                        0      -    18K
-

Now, what do I need to do, so that I get rpool/rb-test
back to where it was when the snapshot rpool/rb-test at 01
was created, but WITHOUT losing what''s at rpool/rb-test at 02?

As there''s no rpool/rb-test at 01, I obviously can''t do a
zfs rollback rpool/rb-test at 01.

I suppose I might be able to do "zfs rollback rpool/rb-test-02 at 01",
but that would destroy rpool/rb-test-02 at 02, wouldn''t it?

Could you please be so kind and show what exactly
needs to be done?

Thanks a lot,
Alexander

On Fri, Dec 11, 2009 at 08:45, Jeff Bonwick <Jeff.Bonwick at sun.com>
wrote:
> Yes, although it''s slightly indirect:
>
>        - make a clone of the snapshot you want to roll back to
>        - promote the clone
>
> See ''zfs promote'' for details.
>
> Jeff
>
> On Fri, Dec 11, 2009 at 08:37:04AM +0100, Alexander Skwar wrote:
> > Hi.
> >
> > Is it possible on Solaris 10 5/09, to rollback to a ZFS snapshot,
> > WITHOUT destroying later created clones or snapshots?
> >
> > Example:
> >
> > --($ ~)-- sudo zfs snapshot rpool/ROOT at 01
> >
> > --($ ~)-- sudo zfs snapshot rpool/ROOT at 02
> >
> > --($ ~)-- sudo zfs clone rpool/ROOT at 02 rpool/ROOT-02
> >
> > --($ ~)-- LC_ALL=C sudo zfs rollback rpool/ROOT at 01
> > cannot rollback to ''rpool/ROOT at 01'': more recent
snapshots exist
> > use ''-r'' to force deletion of the following
snapshots:
> > rpool/ROOT at 02
> >
> > So it isn''t as simple as that. But what needs to be done, to
preserve
> > rpool/ROOT-02?
> >
> > Actually, I''m not concerned (that much) with preserving the
clone
> > rpool/ROOT-02. But I''d like to keep the contents of
rpool/ROOT as
> > it was when I created the @02 snapshot.
> >
> > Is the only possible way to create a backup of rpool/ROOT at 02 (eg.
> > of the snapshot directory /rpool/ROOT/.zfs/snapshots/02) and then
> > restore it later on (eg. backup to tape, backup to someother
filesystem
> > using zfs send|recv, rsync, tar, ...)?
> >
> > Thanks a lot,
> >
> > Alexander
> > --
> > ? Keine Internetzensur in Deutschland! ? http://zensursula.net ? ?
> > ?? ? Lifestream (Twitter, Blog, ?) ??http://alexs77.soup.io/ ? ? ?
> > ? Chat (Jabber/Google Talk) ? a.skwar at gmail.com , AIM: alexws77 ??
> > _______________________________________________
> > zfs-discuss mailing list
> > zfs-discuss at opensolaris.org
> > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
>


-- 

Alexander
-- 
? Keine Internetzensur in Deutschland! ? http://zensursula.net ? ?
?    Lifestream (Twitter, Blog, ?) ? http://alexs77.soup.io/     ?
? Chat (Jabber/Google Talk) ? a.skwar at gmail.com , AIM: alexws77  ?

Sent from Winterthur, ZH, Switzerland
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20091211/7d4b5418/attachment.html>

On Dec 11, 2009, at 4:17 AM, Alexander Skwar <alexanders.mailinglists+nospam
at gmail.com
 > wrote:
> Hello Jeff!
>
> Could you (or anyone else, of course *G*) please show me how?
>
> Situation:
>
> There shall be 2 snapshots of a ZFS called rpool/rb-test
> Let''s call those snapshots "01" and "02".
>
> $ sudo zfs create rpool/rb-test
>
> $ zfs list rpool/rb-test
> NAME            USED  AVAIL  REFER  MOUNTPOINT
> rpool/rb-test    18K   170G    18K  /rpool/rb-test
>
> $ sudo zfs snapshot rpool/rb-test at 01
> $ sudo zfs snapshot rpool/rb-test at 02
> $ sudo zfs clone rpool/rb-test at 02 rpool/rb-test-02
> $ sudo zfs promote rpool/rb-test-02
> $ zfs list -r rpool | grep rb-test
> rpool/rb-test                                              0    
> 170G    18K  /rpool/rb-test
> rpool/rb-test-02                                         18K    
> 170G    18K  /rpool/rb-test-02
> rpool/rb-test-02 at 01                                        0       
> -    18K  -
> rpool/rb-test-02 at 02                                        0       
> -    18K  -
>
> Now, what do I need to do, so that I get rpool/rb-test
> back to where it was when the snapshot rpool/rb-test at 01
> was created, but WITHOUT losing what''s at rpool/rb-test at 02?
>
> As there''s no rpool/rb-test at 01, I obviously can''t do a
> zfs rollback rpool/rb-test at 01.
>
> I suppose I might be able to do "zfs rollback rpool/rb-test-02 at
01",
> but that would destroy rpool/rb-test-02 at 02, wouldn''t it?
>
> Could you please be so kind and show what exactly
> needs to be done?
I think you might be confused at what happens when you promote a  
cloned file system. When it is promoted the clone becomes the file  
system and the original becomes the clone. In this manner you are  
rolling back to the snapshot, but preserving the original as a clone  
and keeping the other snapshots as snapshots of the clone, you just  
might need to fix mountpoints and/or rename file systems to suit your  
taste.

If you "pivot" clones, sub-clones and their snapshots too many times  
you really will make accounting difficult and zfs destroy will take a  
long time to unwind the references, so use sparingly and clean up  
unused references when you can.

-Ross

On Fri, Dec 11, 2009 at 4:17 PM, Alexander Skwar
<alexanders.mailinglists+nospam at gmail.com>
wrote:
> $ sudo zfs create rpool/rb-test
>
> $ zfs list rpool/rb-test
> NAME??????????? USED? AVAIL? REFER? MOUNTPOINT
> rpool/rb-test??? 18K?? 170G??? 18K? /rpool/rb-test
>
> $ sudo zfs snapshot rpool/rb-test at 01
> $ sudo zfs snapshot rpool/rb-test at 02
> $ sudo zfs clone rpool/rb-test at 02 rpool/rb-test-02
at this point it might be better to do some rename and clone

$ sudo zfs rename rpool/rb-test rpool/rb-test-orig
$ sudo zfs clone rpool/rb-test-orig at 01 rpool/rb-test

-- 
Fajar

Hi.

On Fri, Dec 11, 2009 at 15:35, Ross Walker <rswwalker at gmail.com>
wrote:
> On Dec 11, 2009, at 4:17 AM, Alexander Skwar
> <alexanders.mailinglists+nospam at gmail.com> wrote:
>
>> Hello Jeff!
>>
>> Could you (or anyone else, of course *G*) please show me how?
[...]
>> Could you please be so kind and show what exactly
>> needs to be done?
>
> I think you might be confused
Actually, I rather think that I might not have been clear
enough, which then confused you? :)
> at what happens when you promote a cloned file
> system. When it is promoted the clone becomes the file system and the
> original becomes the clone.
Sure. But how does one then rollback to fs at snap-01?
> In this manner you are rolling back to the
> snapshot,
Yes.

Allright, and how do I actually do what I asked for? :)

Please do:

# chose one of the following - or even none, if done as
# root
root_exec=pfexec
root_exec=sudo

$root_exec zfs create rpool/rbtest
$root_exec zfs set mountpoint=/rbtest rpool/rbtest

echo 1 | $root_exec dd of=/rbtest/status
$root_exec zfs snapshot rpool/rbtest at 01

echo 2 | $root_exec dd of=/rbtest/status
$root_exec zfs snapshot rpool/rbtest at 02

$root_exec zfs clone rpool/rbtest at 02 rpool/rbtest-02
$root_exec zfs promote rpool/rbtest-02

--($ ~)-- zfs list | grep rbt
rpool/rbtest
               0   170G  19.5K  /rbtest
rpool/rbtest-02
             52K   170G  19.5K  /rpool/rbtest-02
rpool/rbtest-02 at 01
           16.5K      -  19.5K  -
rpool/rbtest-02 at 02
             16K      -  19.5K  -

Now /rbtest/status contains 2 and so does /rpool/rbtest-02/status:

--($ ~)-- cat /rbtest/status /rpool/rbtest-02/status
2
2

What do I need to do, so /rbtest/status contains 1 (besides
doing something like "echo 1 | sudo dd of=/rbtest/status",
of course *G*) and /rpool/rbtest-02/status (or whatever)
is 2?

It is important that /rbtest/status has exactly that path.

Could you (or anyone? *g*) explain (or better yet: show the
commands)?
> but preserving the original as a clone and keeping the other
> snapshots as snapshots of the clone, you just might need to fix mountpoints
> and/or rename file systems to suit your taste.
How?

--($ ~)-- sudo zfs rename rpool/rbtest-02 at 01 rpool/rbtest at 01
cannot rename to ''rpool/rbtest at 01'': snapshots must be part
of same dataset

I did that, thinking/hoping that this would then allow me to do
a "zfs rollback rpool/rbest at 01".

Alexander
-- 
? Keine Internetzensur in Deutschland! ? http://zensursula.net ? ?
??   Lifestream (Twitter, Blog, ?) ??http://alexs77.soup.io/     ?
? Chat (Jabber/Google Talk) ? a.skwar at gmail.com , AIM: alexws77  ?

Hi!

On Fri, Dec 11, 2009 at 15:55, Fajar A. Nugraha <fajar at fajar.net>
wrote:
> On Fri, Dec 11, 2009 at 4:17 PM, Alexander Skwar
> <alexanders.mailinglists+nospam at gmail.com> wrote:
>> $ sudo zfs create rpool/rb-test
>>
>> $ zfs list rpool/rb-test
>> NAME??????????? USED? AVAIL? REFER? MOUNTPOINT
>> rpool/rb-test??? 18K?? 170G??? 18K? /rpool/rb-test
>>
>> $ sudo zfs snapshot rpool/rb-test at 01
>> $ sudo zfs snapshot rpool/rb-test at 02
>> $ sudo zfs clone rpool/rb-test at 02 rpool/rb-test-02
>
> at this point it might be better to do some rename and clone
>
> $ sudo zfs rename rpool/rb-test rpool/rb-test-orig
> $ sudo zfs clone rpool/rb-test-orig at 01 rpool/rb-test
AH!

Now I understand!

The crucial point here is the "zfs rename", which
"moves" the original stuff out of the way and then
whatever data is required (eg snap @01 or @02)
is "moved" to the correct spot by cloning the wanted
snapshot.

That''s quite clever - thanks a lot! It''s clear now.

But this has the "chance" of becoming rather confusing
if too many snapshots or "branches" are used. But
that''s the way it is? :-)

Thanks again.

Cheers,

Alexander
-- 
? Keine Internetzensur in Deutschland! ? http://zensursula.net ? ?
??   Lifestream (Twitter, Blog, ?) ??http://alexs77.soup.io/     ?
? Chat (Jabber/Google Talk) ? a.skwar at gmail.com , AIM: alexws77  ?

Sent from Winterthur, ZH, Switzerland

On Dec 11, 2009, at 3:26 PM, Alexander Skwar <alexanders.mailinglists+nospam
at gmail.com
 > wrote:
> Hi!
>
> On Fri, Dec 11, 2009 at 15:55, Fajar A. Nugraha <fajar at fajar.net>
> wrote:
>> On Fri, Dec 11, 2009 at 4:17 PM, Alexander Skwar
>> <alexanders.mailinglists+nospam at gmail.com> wrote:
>>> $ sudo zfs create rpool/rb-test
>>>
>>> $ zfs list rpool/rb-test
>>> NAME            USED  AVAIL  REFER  MOUNTPOINT
>>> rpool/rb-test    18K   170G    18K  /rpool/rb-test
>>>
>>> $ sudo zfs snapshot rpool/rb-test at 01
>>> $ sudo zfs snapshot rpool/rb-test at 02
>>> $ sudo zfs clone rpool/rb-test at 02 rpool/rb-test-02
>>
>> at this point it might be better to do some rename and clone
>>
>> $ sudo zfs rename rpool/rb-test rpool/rb-test-orig
>> $ sudo zfs clone rpool/rb-test-orig at 01 rpool/rb-test
>
> AH!
>
> Now I understand!
>
> The crucial point here is the "zfs rename", which
> "moves" the original stuff out of the way and then
> whatever data is required (eg snap @01 or @02)
> is "moved" to the correct spot by cloning the wanted
> snapshot.
>
> That''s quite clever - thanks a lot! It''s clear now.
>
> But this has the "chance" of becoming rather confusing
> if too many snapshots or "branches" are used. But
> that''s the way it is? :-)
Juliet:
"What''s in a name? That which we call a rose
By any other name would smell as sweet."

Romeo and Juliet (II, ii, 1-2)

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20091211/0497eebc/attachment.html>