parted behavior change between 3.1 and 3.2

parted behavior change between 3.1 and 3.2

[Loic Dachary]

Hi Philip,

In the context of the Ceph ( http://ceph.com/ ) project, I'm trying to figure out why parted 3.2 fixes a problem that shows in 3.1 [1]. In a nutshell, calling partprobe sometime fails with "device busy" with 3.1 and when using 3.2 that never happens (all things are the same, only parted changes). I took a look at the commits in http://git.savannah.gnu.org/cgit/parted.git/log/ but I didn't see a commit that would explain such a behavior change.

Any pointer or idea you may have would be greatly appreciated :-)

Cheers

[1] partprobe intermittent issues during ceph-disk prepare http://tracker.ceph.com/issues/15176
-- 
Loïc Dachary, Artisan Logiciel Libre
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: parted behavior change between 3.1 and 3.2

[Phil Susi]

On 5/23/2016 3:41 AM, Loic Dachary wrote:
> Hi Philip,
> 
> In the context of the Ceph ( http://ceph.com/ ) project, I'm trying
> to figure out why parted 3.2 fixes a problem that shows in 3.1 [1].
> In a nutshell, calling partprobe sometime fails with "device busy"
> with 3.1 and when using 3.2 that never happens (all things are the
> same, only parted changes). I took a look at the commits in
> http://git.savannah.gnu.org/cgit/parted.git/log/ but I didn't see a
> commit that would explain such a behavior change.
> 
> Any pointer or idea you may have would be greatly appreciated :-)

An strace would help, but if I had to guess, one significant change
between 3.1 and 3.2 was the addition of the ability to resize mounted
partitions, so it could be that partprobe is adjusting the size of a
changed partition, and it used to not be able to do that if it is mounted.

Re: parted behavior change between 3.1 and 3.2

[Loic Dachary]

Hi Phil,

On 24/05/2016 19:16, Phil Susi wrote:
> On 5/23/2016 3:41 AM, Loic Dachary wrote:
>> Hi Philip,
>>
>> In the context of the Ceph ( http://ceph.com/ ) project, I'm trying
>> to figure out why parted 3.2 fixes a problem that shows in 3.1 [1].
>> In a nutshell, calling partprobe sometime fails with "device busy"
>> with 3.1 and when using 3.2 that never happens (all things are the
>> same, only parted changes). I took a look at the commits in
>> http://git.savannah.gnu.org/cgit/parted.git/log/ but I didn't see a
>> commit that would explain such a behavior change.
>>
>> Any pointer or idea you may have would be greatly appreciated :-)
> 
> An strace would help, but if I had to guess, one significant change
> between 3.1 and 3.2 was the addition of the ability to resize mounted
> partitions, so it could be that partprobe is adjusting the size of a
> changed partition, and it used to not be able to do that if it is mounted.
> 

That's an interesting lead, I'll double check if that could be the cause. Thanks very much for the hint !

I'll let you know how it evolves.

Cheers

-- 
Loïc Dachary, Artisan Logiciel Libre
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

parted behavior change between 3.1 and 3.2

[Loic Dachary]

Hi Brian,

In the context of the Ceph ( http://ceph.com/ ) project, I'm trying to figure out why parted 3.2 fixes a problem that shows in 3.1 [1]. In a nutshell, calling partprobe sometime fails with "device busy" with 3.1 and when using 3.2 that never happens (all things are the same, only parted changes). I took a look at the commits in http://git.savannah.gnu.org/cgit/parted.git/log/ but I didn't see a commit that would explain such a behavior change.

Any pointer or idea you may have would be greatly appreciated :-)

Cheers

[1] partprobe intermittent issues during ceph-disk prepare http://tracker.ceph.com/issues/15176
-- 
Loïc Dachary, Artisan Logiciel Libre
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html