Re: [GIT PULL] drbd-8.4 for mainline

[Jens Axboe <axboe@kernel.dk>]

On 2011-08-24 18:00, Kyle Moffett wrote:
> On Wed, Aug 24, 2011 at 10:41, Philipp Reisner
> <philipp.reisner@linbit.com> wrote:
>> Hi Jens,
>>
>> First the announcement of drbd-8.4, then the git pull-request text:
>>
>> We are proud to announce the availability of DRBD-8.4.0.
>>
>>  The most noticeable change is the support for multiple replicated
>>  volumes in a single DRBD connection.
>>  Write-ordering is obeyed among all writes in all volumes in a
>>  single connection.
>>  This feature is really important for users who DRBD for mirroring
>>  over longer distances. (Protocol A).
>>
>> We do not only release DRBD-8.4.0 today:
>> The DRBD User's Guide was reviewed and updated to match DRBD-8.4.
>>
>> I suggest to everybody who considers to upgrade from 8.3 to 8.4
>> to have a look at the "Recent changes" appendix of the UG:
>> http://www.drbd.org/users-guide/ap-recent-changes.html
>>
>>  This release brings a new meta-data format. Forward (8.3 -> 8.4)
>>  conversion happens complete seamless. Backward conversion
>>  is done by a single command (drbdadm apply-al res).
>>
>>  This release is protocol compatible with all it predecessor.
>>  Although, we do not recommend to run it in 8.3 - 8.4 for long
>>  time frames. We recommend to use that capability only for the
>>  rolling upgrade.
>>
>>  drbdadm of 8.4 can parse config files of 8.3. We recommend
>>  to switch to the new configuration syntax after the upgrade
>>  of both nodes. (Use drbdadm dump to learn about the new
>>  config syntax)
> 
> Hm...
> 
> That's a lot of patches (including some protocol changes) that have not
> yet been reviewed by other kernel developers.
> 
> By officially releasing the kernel and user-space bits and then posting
> them to LKML and expecting them to be merged as-is, you are not really
> following the linux kernel development process.
> 
> Some of the reverts and commit messages make me concerned that your
> patch series has bisection issues; are you sure it compiles and runs
> after every patch?
> 
> I'm obviously not anywhere in the maintenance chain for this code, but
> it does look really funny.

That was my exact response a few weeks back, but I don't recall seeing
anything until this email today. Philipp, has this been reviewed at all
outside your internal group?

-- 
Jens Axboe