Re: Modifying RBD image changes it's snapshot

[Josh Durgin <josh.durgin@inktank.com>]

On 06/26/2013 11:15 AM, Josh Durgin wrote:
> On 06/26/2013 05:40 AM, Karol Jurak wrote:
>> Hi,
>>
>> I'm using ceph 0.56.6 and kernel 3.9.7 and it looks like modifying RBD
>> image
>> also changes it's snapshot. I can reproduce this as follows:
>>
>> # create and map an image
>> rbd create --size 128 test-1
>> rbd map test-1
>>
>> # write some data to the image
>> dd if=/dev/zero of=/dev/rbd/rbd/test-1 bs=1M count=128
>>
>> # create and map a snapshot
>> rbd snap create test-1@snap-1
>> rbd map test-1@snap-1
>>
>> # verify that image and snapshot are identical
>> md5sum /dev/rbd/rbd/test-1 /dev/rbd/rbd/test-1@snap-1
>> fde9e0818281836e4fc0edfede2b8762  /dev/rbd/rbd/test-1
>> fde9e0818281836e4fc0edfede2b8762  /dev/rbd/rbd/test-1@snap-1
>>
>> # modify the image
>> dd if=/dev/urandom of=/dev/rbd/rbd/test-1 bs=512 count=1
>>
>> # compare checksums again
>> md5sum /dev/rbd/rbd/test-1 /dev/rbd/rbd/test-1@snap-1
>> 1d942c8a5bc7480cecb945ea0d020eed  /dev/rbd/rbd/test-1
>> 1d942c8a5bc7480cecb945ea0d020eed  /dev/rbd/rbd/test-1@snap-1
>>
>> Checksums are identical although the snapshot isn't supposed to be
>> modified.
>>
>
> Just reproduced on 3.10-rc7 as well. It seems the snapshot context
> loading is broken for format 1 a t least, since unmapping and mapping
> after the snapshot exists still has the same problem. I added
> http://tracker.ceph.com/issues/5464 to track this.
>
> Apparently the regression test for this hasn't been running, or we
> would've caught this sooner.

There's a fix for this on the stable kernel in the wip-snapc-3.9.y
branch of ceph-client.git.

Josh