Re: Question about reflink feature

Re: Question about reflink feature

[Darrick J. Wong]

[add linux-xfs to cc]

On Wed, Jul 25, 2018 at 03:48:19PM +0800, Xiao Yang wrote:
> Hi Darrick,
> 
> Sorry to bother you.
> 
> Currently, all reflink tests for XFS in xfstests are skipped even if we
> use the lastest kernel and xfsprogs.
> It seems that mkfs.xfs doesn't enable reflink by default, do you want to
> enable it by default in later
> xfsprogs version?

It will be enabled by default when we've had a long enough period to
test it and make sure it's stable for general use.  That may happen some
time in early 2019.

> Shoud we enable it forcely by "mkfs.xfs -m reflink" in
> xfstests if mkfs.xfs supports it?

For now, yes.

--D

> 
> Thanks,
> Xiao Yang
> 
> 

Re: Question about reflink feature

[Xiao Yang]

Hi Darrick,

Thanks for your explanation.

rmapbt, reflink and dedupe features will be enabled by default together?

Thanks
Xiao Yang
On 2018/07/26 3:16, Darrick J. Wong wrote:
> [add linux-xfs to cc]
>
> On Wed, Jul 25, 2018 at 03:48:19PM +0800, Xiao Yang wrote:
>> Hi Darrick,
>>
>> Sorry to bother you.
>>
>> Currently, all reflink tests for XFS in xfstests are skipped even if we
>> use the lastest kernel and xfsprogs.
>> It seems that mkfs.xfs doesn't enable reflink by default, do you want to
>> enable it by default in later
>> xfsprogs version?
> It will be enabled by default when we've had a long enough period to
> test it and make sure it's stable for general use.  That may happen some
> time in early 2019.
>
>> Shoud we enable it forcely by "mkfs.xfs -m reflink" in
>> xfstests if mkfs.xfs supports it?
> For now, yes.
>
> --D
>
>> Thanks,
>> Xiao Yang
>>
>>
>
>

Re: Question about reflink feature

[Eric Sandeen]

On 7/26/18 6:34 PM, Xiao Yang wrote:
> Hi Darrick,
> 
> Thanks for your explanation.
> 
> rmapbt, reflink and dedupe features will be enabled by default together?

dedup is not a "feature" of xfs by itself; rather there are userspace
applications which can use clone_file_range (related to reflink) to
do out of band deduplication from userspace.

I believe that rmapbt has a fair bit of performance overhead, and may
not get enabled by default.  Not certain about this though.

-Eric

> Thanks
> Xiao Yang
> On 2018/07/26 3:16, Darrick J. Wong wrote:
>> [add linux-xfs to cc]
>>
>> On Wed, Jul 25, 2018 at 03:48:19PM +0800, Xiao Yang wrote:
>>> Hi Darrick,
>>>
>>> Sorry to bother you.
>>>
>>> Currently, all reflink tests for XFS in xfstests are skipped even if we
>>> use the lastest kernel and xfsprogs.
>>> It seems that mkfs.xfs doesn't enable reflink by default, do you want to
>>> enable it by default in later
>>> xfsprogs version?
>> It will be enabled by default when we've had a long enough period to
>> test it and make sure it's stable for general use.  That may happen some
>> time in early 2019.
>>
>>> Shoud we enable it forcely by "mkfs.xfs -m reflink" in
>>> xfstests if mkfs.xfs supports it?
>> For now, yes.
>>
>> --D
>>
>>> Thanks,
>>> Xiao Yang
>>>
>>>
>>
>>
> 
> 
> 
> -- 
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>