Thanks Hequn. +1 for starting RC of 1.9.2. With the development of our 1.10 bug fix, we have fixed 137 issues for 1.9.2. It's worth releasing 1.9.2, which really solves a lot of urgent bugs for users.
[1] https://issues.apache.org/jira/browse/FLINK-15726?jql=project%20%3D%20FLINK%20AND%20fixVersion%20%3D%201.9.2 Best, Jingsong Lee On Thu, Jan 23, 2020 at 3:06 AM Bowen Li <bowenl...@gmail.com> wrote: > +1. Thanks Hequn > > On Wed, Jan 22, 2020 at 8:39 AM Till Rohrmann <trohrm...@apache.org> > wrote: > > > Thanks for resuming the discussion Hequn. +1 for starting with the RC > > creation. Thanks for driving the release process! > > > > Cheers, > > Till > > > > On Tue, Jan 21, 2020 at 11:02 PM jincheng sun <sunjincheng...@gmail.com> > > wrote: > > > > > Cool, looking forward the first RC of release 1.9.2. > > > > > > Best, > > > Jincheng > > > > > > > > > Hequn Cheng <chenghe...@gmail.com> 于2020年1月21日周二 下午10:02写道: > > > > > > > Hi everyone, > > > > > > > > Considering that we are on the finishing line for release-1.10.0 and > > that > > > > there are no blockers for release-1.9.2, I'm proposing to resume this > > > > discussion. > > > > If there are no more concerns or any critical issues for releasing > > > 1.9.2, I > > > > would like to create the first RC soon :) > > > > > > > > Best, > > > > Hequn > > > > > > > > On Wed, Dec 18, 2019 at 7:10 PM Hequn Cheng <chenghe...@gmail.com> > > > wrote: > > > > > > > > > Hi Till, > > > > > > > > > > I agree with your concerns and thanks a lot for your feedback! > > > > > > > > > > In fact, I also have those concerns. The reasons I started the > > DISCUSS > > > > are: > > > > > - Considering the low capacities, I want to start the discussion > > > earlier > > > > > so that we can have more time to collect information and planning > for > > > the > > > > > release. Usually, it takes a couple of weeks to address the > blockers. > > > > > - The blockers of 1.9.2 are probably also blockers for 1.10.0, take > > > > > FLINK-15266 as an example. I can help to drive to solve the problem > > so > > > > that > > > > > 1.10.0 can also benefit from it. > > > > > > > > > > However, to ease the load, we can vote after Christmas if there are > > no > > > > > more blockers at that time. > > > > > There is a chance that the 1.10 blocker resolution has calmed down > a > > > bit. > > > > > > > > > > Best, > > > > > Hequn > > > > > > > > > > On Wed, Dec 18, 2019 at 7:05 PM Hequn Cheng <chenghe...@gmail.com> > > > > wrote: > > > > > > > > > >> Hi Jincheng, > > > > >> > > > > >> Yes, your help would be very helpful. Thanks a lot! > > > > >> > > > > >> Best, Hequn > > > > >> > > > > >> On Wed, Dec 18, 2019 at 4:52 PM Till Rohrmann < > trohrm...@apache.org > > > > > > > >> wrote: > > > > >> > > > > >>> Hi Hequn, > > > > >>> > > > > >>> thanks for starting this discussion. In general I think it is a > > good > > > > idea > > > > >>> to release often. Hence, I also believe it is time for another > bug > > > fix > > > > >>> release for 1.9. > > > > >>> > > > > >>> The thing I'm wondering is whether we are stretching our > resources > > a > > > > bit > > > > >>> too much if we now start with a 1.9.2 release vote because of the > > > > ongoing > > > > >>> 1.10 release testing work. My concern is that we don't have the > > > > >>> capacities > > > > >>> to properly test the 1.9 RC. Additionally, many people will be on > > > > >>> Christmas > > > > >>> vacations during the next 2 weeks. If this is it not a problem > and > > we > > > > >>> have > > > > >>> enough resources for a parallel release then we can start it. > > > Otherwise > > > > >>> I'd > > > > >>> say that we wait until the 1.10 blocker resolution has calmed > down > > a > > > > bit. > > > > >>> > > > > >>> Cheers, > > > > >>> Till > > > > >>> > > > > >>> On Wed, Dec 18, 2019 at 8:47 AM jincheng sun < > > > sunjincheng...@gmail.com > > > > > > > > > >>> wrote: > > > > >>> > > > > >>> > Thanks for bring up the discussion Hequn. I would like to give > > you > > > a > > > > >>> hand > > > > >>> > at the last stage when the RC is finished.(If you need) :) > > > > >>> > > > > > >>> > Best, > > > > >>> > Jincheng > > > > >>> > > > > > >>> > Best, > > > > >>> > Jincheng > > > > >>> > @sunjincheng121 <https://twitter.com/sunjincheng121> > > > > >>> > > > > > >>> > > > > > >>> > Hequn Cheng <chenghe...@gmail.com> 于2019年12月18日周三 下午3:44写道: > > > > >>> > > > > > >>> > > Hi everyone, > > > > >>> > > > > > > >>> > > It has already been two months since we released the Flink > > 1.9.1. > > > > >>> > > We already have many important bug fixes from which our users > > can > > > > >>> benefit > > > > >>> > > in the release-1.9 branch (85 resolved issues). > > > > >>> > > Therefore, I propose to create the next bug fix release for > > Flink > > > > >>> 1.9. > > > > >>> > > > > > > >>> > > Most notable fixes are: > > > > >>> > > > > > > >>> > > - [FLINK-14074] MesosResourceManager can't create new > > > taskmanagers > > > > in > > > > >>> > > Session Cluster Mode. > > > > >>> > > - [FLINK-14995] Kinesis NOTICE is incorrect > > > > >>> > > - [FLINK-15013] Flink (on YARN) sometimes needs too many > slots > > > > >>> > > - [FLINK-15036] Container startup error will be handled out > > side > > > of > > > > >>> the > > > > >>> > > YarnResourceManager's main thread > > > > >>> > > - [FLINK-14315] NPE with JobMaster.disconnectTaskManager > > > > >>> > > > > > > >>> > > Furthermore, there is one issue marked as blocker for 1.9.2 > > which > > > > >>> should > > > > >>> > be > > > > >>> > > merged before 1.9.2 release: > > > > >>> > > > > > > >>> > > - [FLINK-15266] NPE in blink planner code gen (reviewing) > > > > >>> > > > > > > >>> > > If there are any other blocker issues need to be fixed in > > 1.9.2, > > > > >>> please > > > > >>> > let > > > > >>> > > me know. > > > > >>> > > I will kick off the release process once blocker issues have > > been > > > > >>> merged. > > > > >>> > > > > > > >>> > > It would be very appreciated if there is any PMC can help > with > > > the > > > > >>> final > > > > >>> > > steps of the release process. > > > > >>> > > > > > > >>> > > Best, > > > > >>> > > Hequn > > > > >>> > > > > > > >>> > > > > > >>> > > > > >> > > > > > > > > > > -- Best, Jingsong Lee