From mboxrd@z Thu Jan 1 00:00:00 1970 From: Loic Dachary Subject: Re: rbd-mirror related jewel backports Date: Fri, 12 Aug 2016 11:12:01 +0200 Message-ID: <57AD92E1.9050409@dachary.org> References: <57AAEEC2.3060108@dachary.org> <57AB95F9.3050101@dachary.org> <57AC847F.3020601@dachary.org> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Return-path: Received: from relay5-d.mail.gandi.net ([217.70.183.197]:39144 "EHLO relay5-d.mail.gandi.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751356AbcHLJMH (ORCPT ); Fri, 12 Aug 2016 05:12:07 -0400 In-Reply-To: Sender: ceph-devel-owner@vger.kernel.org List-ID: To: dillaman@redhat.com Cc: Mykola Golub , Ceph Development Hi Jason, On 11/08/2016 17:01, Jason Dillaman wrote: > OK, that's fine with me. The only ones left on your original list are > (in order): Excellent, that makes our work a lot easier :-) I'll start testing https://github.com/ceph/ceph/pull/10678 and once merged it will clear 6 backports. > > http://tracker.ceph.com/issues/16902 (master 10484) (Tue, 2 Aug 2016 > 20:14:54 +0300) > http://tracker.ceph.com/issues/16950 (master 10574) (Sun, 7 Aug 2016 > 13:33:45 +0300) > http://tracker.ceph.com/issues/16978 (master 10613) (Tue, 9 Aug 2016 > 16:18:49 +0300) > > On Thu, Aug 11, 2016 at 9:58 AM, Loic Dachary wrote: >> I can try to do that, it is simple enough. Because I'm not very familiar with the code base, it is entirely possible that I end up cherry-picking a lot more than what is reasonable when doing that. It is sometime non trivial to figure out when to stop. Who knows: I may end up cherry-picking *everything* from master :-) I'll call for help if it drifts too much. >> >> Cheers >> >> On 11/08/2016 15:22, Jason Dillaman wrote: >>> No worries. >>> >>> My usual routine just involves pulling up master and jewel branch >>> merge logs for all RBD-related subdirectories in two terminals (for PR >>> merge timeline reference), pulling up the original PR to backport, and >>> then just cherry-pick until a conflict is hit. When that happens, I >>> locate which previous PR made the conflicting change and cherry-pick >>> it into the same backport branch. >>> >>> With this technique, you'll end up with PRs with the same >>> cherry-picked commits. You'll be able to cleanly merge them into a >>> testing branch, but when it comes time to merge into the jewel branch >>> you'll need to rebase some PRs to automatically strip out the >>> duplicate cherry-picks. >>> >>> On Wed, Aug 10, 2016 at 5:00 PM, Loic Dachary wrote: >>>> >>>> >>>> On 10/08/2016 22:14, Jason Dillaman wrote: >>>>> Loic, >>>>> >>>>> If you want, I can create the backport PRs for those tickets. >>>> >>>> That would be great. Alternatively, if that's less time consuming for you, just suggest a backport order and I'll figure it out. >>>> >>>> Cheers >>>> >>>>> >>>>> >>>>> On Wed, Aug 10, 2016 at 5:07 AM, Loic Dachary wrote: >>>>>> Hi Mykola, >>>>>> >>>>>> There are seven rbd-mirror related backports which do not cherry-pick cleanly. I've not looked at each of them yet and I would very much appreciate your advice about the best way to proceed. >>>>>> >>>>>> http://tracker.ceph.com/issues/16511 >>>>>> http://tracker.ceph.com/issues/16512 >>>>>> http://tracker.ceph.com/issues/16658 >>>>>> http://tracker.ceph.com/issues/16701 >>>>>> http://tracker.ceph.com/issues/16747 >>>>>> http://tracker.ceph.com/issues/16902 >>>>>> http://tracker.ceph.com/issues/16978 >>>>>> >>>>>> Maybe they should be cherry-picked in a specific order in order to avoid conflicts ? For instance, maybe one of the pending backports that you can see at >>>>>> >>>>>> https://github.com/ceph/ceph/pulls?q=is%3Aopen+is%3Apr+milestone%3Ajewel+label%3Arbd >>>>>> >>>>>> would avoid some of the conflicts if merged ? I'm thinking specifically >>>>>> >>>>>> https://github.com/ceph/ceph/pull/10646 >>>>>> https://github.com/ceph/ceph/pull/10644 >>>>>> >>>>>> What I'm looking for, ideally, is a way to order the cherry-picks to avoid any conflict. That may involve backporting a few additional pull requests (for instance when the type of an argument changes from int64_t to uint64_t, that creates a lot of potential conflicts and backporting that change could help, even if it is not required to fix a given bug. >>>>>> >>>>>> There are three backports that also do not apply because of a conflict but do not seem to be related to rbd-mirror. Unless you advise otherwise, I'll attempt to resolve these individually and assume there is complex interaction with other backports. >>>>>> >>>>>> http://tracker.ceph.com/issues/16950 >>>>>> http://tracker.ceph.com/issues/16904 >>>>>> http://tracker.ceph.com/issues/16735 >>>>>> >>>>>> Thanks in advance for your guidance ! >>>>>> >>>>>> -- >>>>>> 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 >>>>> >>>>> >>>>> >>>> >>>> -- >>>> Loïc Dachary, Artisan Logiciel Libre >>> >>> >>> >> >> -- >> Loïc Dachary, Artisan Logiciel Libre > > > -- Loïc Dachary, Artisan Logiciel Libre