The underlying issue has existed forever AFAIK, but the issue has been
masked since Infernalis due to the deep-flatten changes. If you were to use
krbd or an older librbd client to write to the image, the same issue would
appear (until the PR is merged and backported).
On Nov 30, 2016, at 1:00 AM,
Thanks for the quick confirm and fix.
I reviewed the pull request. Seems that this fix is going into Jewel and
Hammer.
So does this mean this issue also exists in Jewel? why I test Jewel and
cannot reproduce this issue??
Zhongyan
On Wed, Nov 30, 2016 at 12:32 PM, Venky Shankar wrote:
> On 16-11
On 16-11-30 12:07:33, Zhongyan Gu wrote:
> Jason,
> I test Jewel and confirmed Jewel has no such issue.
> Could you tell me what is the specific pull that can be backported to
> hammer to fix this issue??
PR: https://github.com/ceph/ceph/pull/12218
>
> Zhongyan
>
> On Tue, Nov 29, 2016 at 10:52
Jason,
I test Jewel and confirmed Jewel has no such issue.
Could you tell me what is the specific pull that can be backported to
hammer to fix this issue??
Zhongyan
On Tue, Nov 29, 2016 at 10:52 PM, Jason Dillaman
wrote:
> You are correct that there is an issue with the Hammer version when
> ca
You are correct that there is an issue with the Hammer version when
calculating diffs. If the clone has an object that obscures an extent
within the parent image but didn't exist within the first snapshot of
the clone, the diff results from the parent image won't be included in
the result. I'll ope
Thank you Jason.
We are designing a backup system for production cluster based on ceph's
export /import diff feature.
We found this issue and hopefully it can be confirmed and then fixed soon.
If you need any more information for debugging, Please just let me know.
Thanks,
Zhongyan
On Mon, Nov 28
OK, in that case, none of my previous explanation is relevant. I'll
spin up a hammer cluster and try to reproduce.
On Wed, Nov 23, 2016 at 9:13 PM, Zhongyan Gu wrote:
> BTW, I used Hammer 0.94.5 to do the test.
>
> Zhongyan
>
> On Thu, Nov 24, 2016 at 10:07 AM, Zhongyan Gu wrote:
>>
>> Thank you
BTW, I used Hammer 0.94.5 to do the test.
Zhongyan
On Thu, Nov 24, 2016 at 10:07 AM, Zhongyan Gu wrote:
> Thank you Jason. My test shows in the following case, image B will be
> exactly same:
> 1. clone image A from parent:
> #rbd clone 1124-parent@snap1 A
>
> 2. create snap for A
> #rbd snap c
Thank you Jason. My test shows in the following case, image B will be
exactly same:
1. clone image A from parent:
#rbd clone 1124-parent@snap1 A
2. create snap for A
#rbd snap create A@snap1
3. create empty image B
#rbd create B -s 1
4. export-diff A then impor-diff B:
#rbd export-diff A@snap1 -
What you are seeing sounds like a side-effect of deep-flatten support.
If you write to an unallocated extent within a cloned image, the
associated object extent must be read from the parent image, modified,
and written to the clone image.
Since the Infernalis release, this process has been tweaked
Let me make the issue more clear.
Suppose I cloned image A from a parent image and create snap1 for image A
and then make some change of image A.
If I did the rbd export-diff @snap1. how should I prepare the existing
image B to make sure it will be exactly same with image A@snap1 after
import-dif
Thanks Jason, very clear explanation.
However, I found some strange behavior when export-diff on a cloned image,
not sure it is a bug on calc_snap_set_diff().
The test is,
Image A is cloned from a parent image. then create snap1 for image A.
The content of export-diff A@snap1 will be changed when u
On Tue, Nov 22, 2016 at 5:31 AM, Zhongyan Gu wrote:
> So if initial snapshot is NOT specified, then:
> rbd export-diff image@snap1 will diff all data to snap1. this cmd equals to
> :
> rbd export image@snap1. Is my understand right or not??
While they will both export all data associated w/ imag
Hi there,
According to the official man page:
http://docs.ceph.com/docs/jewel/man/8/rbd/
export-diff [–from-snap snap-name] [–whole-object] (image-spec | snap-spec)
dest-path
Exports an incremental diff for an image to dest path (use - for stdout).
If an initial snapshot is specified, only changes
14 matches
Mail list logo
