Revert system state to snapshot

Revert system state to snapshot

[Yevhen Kyriukha]

Hi!

I'm wondering how can revert to snapshot by mounting file system as
writable in this state.

I know that NILFS2 currently doesn't support this feature, but I think
that implementation of such functionality is highly desirable.
Is there any date or estimates when this feature will be implemented?

Best regards,
Yevhen
--
To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Revert system state to snapshot

[Vyacheslav Dubeyko]

Hi Yevhen,

On Fri, 2013-10-04 at 15:28 +0300, Yevhen Kyriukha wrote:
> Hi!
> 
> I'm wondering how can revert to snapshot by mounting file system as
> writable in this state.
> 
> I know that NILFS2 currently doesn't support this feature, but I think
> that implementation of such functionality is highly desirable.
> Is there any date or estimates when this feature will be implemented?
> 

If you would like to implement this feature then it will be great
contribution to the NILFS2 project. You are welcome.

Also it can be valuable your vision of this feature. Could you describe
what you expect from this feature and how you want to use it? A user's
vision is different from developer's vision very frequently.

As far as I can judge, we haven't any implementation activity in this
direction. If I am not fully informed then Ryusuke corrects me.

Thanks,
Vyacheslav Dubeyko.


--
To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Revert system state to snapshot

[Сергей Александров]

Hello,

Personally for me, it will be great to an utility like this(supporting
the current utility conventions):
rvcp cp | ss [device] checkpoint-number
This should only work on an unmounted partition(else error), cp type
should be the one set by user (sp | ss),
The question should be asked that the operation will make all the data
written after this cp (cp date here) will be unrecoverably lost
continue any way?[yes/no], requiring full answer.
After this the FS should be in the state described by this cp and can
be mounted rw. All cps after given should be eliminated including ss
cps.

The use case is reverting system state(even the rootfs with the help
of live cd or live usb) to the previously made ss or existing cp
without any need of having another hdd, coping data from ro mounted cp
to the drive and than copying it back. This should ease FS modifying
related experiments greatly.

Thanks,
Aleksandrov Sergey

--------------------------------------------------
Александров Сергей Васильевич


2013/10/4 Vyacheslav Dubeyko <slava-yeENwD64cLxBDgjK7y7TUQ@public.gmane.org>:
> Hi Yevhen,
>
> On Fri, 2013-10-04 at 15:28 +0300, Yevhen Kyriukha wrote:
>> Hi!
>>
>> I'm wondering how can revert to snapshot by mounting file system as
>> writable in this state.
>>
>> I know that NILFS2 currently doesn't support this feature, but I think
>> that implementation of such functionality is highly desirable.
>> Is there any date or estimates when this feature will be implemented?
>>
>
> If you would like to implement this feature then it will be great
> contribution to the NILFS2 project. You are welcome.
>
> Also it can be valuable your vision of this feature. Could you describe
> what you expect from this feature and how you want to use it? A user's
> vision is different from developer's vision very frequently.
>
> As far as I can judge, we haven't any implementation activity in this
> direction. If I am not fully informed then Ryusuke corrects me.
>
> Thanks,
> Vyacheslav Dubeyko.
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Revert system state to snapshot

[Yevhen Kyriukha]

I think that Sergey's suggestion is great to start with.
Such kind of tool will be the easiest to implement.

The ideal (future) solution I see here is to implement "revert to
snapshot" as VCS implements "revert to commit": when you revert to
specific revision and make some changes then this changes can be
committed in new revision.
In such a way you do not need to change existing data and just append changes.

Best regards,
Yevhen

2013/10/4 Сергей Александров <splavgm@gmail.com>:
> Hello,
>
> Personally for me, it will be great to an utility like this(supporting
> the current utility conventions):
> rvcp cp | ss [device] checkpoint-number
> This should only work on an unmounted partition(else error), cp type
> should be the one set by user (sp | ss),
> The question should be asked that the operation will make all the data
> written after this cp (cp date here) will be unrecoverably lost
> continue any way?[yes/no], requiring full answer.
> After this the FS should be in the state described by this cp and can
> be mounted rw. All cps after given should be eliminated including ss
> cps.
>
> The use case is reverting system state(even the rootfs with the help
> of live cd or live usb) to the previously made ss or existing cp
> without any need of having another hdd, coping data from ro mounted cp
> to the drive and than copying it back. This should ease FS modifying
> related experiments greatly.
>
> Thanks,
> Aleksandrov Sergey
>
> --------------------------------------------------
> Александров Сергей Васильевич
>
>
> 2013/10/4 Vyacheslav Dubeyko <slava-yeENwD64cLxBDgjK7y7TUQ@public.gmane.org>:
>> Hi Yevhen,
>>
>> On Fri, 2013-10-04 at 15:28 +0300, Yevhen Kyriukha wrote:
>>> Hi!
>>>
>>> I'm wondering how can revert to snapshot by mounting file system as
>>> writable in this state.
>>>
>>> I know that NILFS2 currently doesn't support this feature, but I think
>>> that implementation of such functionality is highly desirable.
>>> Is there any date or estimates when this feature will be implemented?
>>>
>>
>> If you would like to implement this feature then it will be great
>> contribution to the NILFS2 project. You are welcome.
>>
>> Also it can be valuable your vision of this feature. Could you describe
>> what you expect from this feature and how you want to use it? A user's
>> vision is different from developer's vision very frequently.
>>
>> As far as I can judge, we haven't any implementation activity in this
>> direction. If I am not fully informed then Ryusuke corrects me.
>>
>> Thanks,
>> Vyacheslav Dubeyko.
>>
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in
>> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html