Thanks Yu for this proposal. +1(non-binding) for the overall design. I have 2 questions about the result check though. Posted in the discussion ML thread.
Thanks, Zhu Zhu Becket Qin <becket....@gmail.com> 于2019年11月21日周四 上午11:57写道： > Hi Yu, > > Thanks for updating the FLIP wiki. +1 from me. I do not have further > questions. > > Jiangjie (Becket) Qin > > > > On Wed, Nov 20, 2019 at 8:58 PM Yu Li <car...@gmail.com> wrote: > > > Thanks all for the vote and further discussions. > > > > I have updated the FLIP document according to aihua's reply (with some > > minor refinement/supplements). For your convenience, below are the > contents > > I added: > > *====================================* > > > > *There're also other dimensions other than Flink characteristics, > > including:* > > > > - *Record size: to check both the processing (records/s) and data > > (bytes/s) throughput, we will test the 10B, 100B and 1KB record size > for > > each test job.* > > - *Resource for each task: we will use the Flink default settings to > > cover the most used cases.* > > - *Job Parallelism: we will increase the parallelism to saturate the > > system until back-pressure.* > > - *Source and Sink: to focus on Flink performance, we generate the > > source data randomly and use a blackhole consumer as the sink.* > > > > *====================================* > > > > @becket.qin <becket....@gmail.com> please let us know if the updates > look > > good to you to turn your voting to a complete binding +1, or if any more > > concerns/comments. Thanks! > > > > @voters: please also take a look at the updates and let us know if any > new > > comments. Thanks. > > > > Best Regards, > > Yu > > > > > > On Wed, 20 Nov 2019 at 11:10, aihua li <liaihua1...@gmail.com> wrote: > > > >> hi，Becket, > >> > >> Thanks for the comments! > >> > >> > 1. How do the testing records look like? The size and key > distributions. > >> > >> The records looks like a long string,which default size is 1k. The key > is > >> randomly generated according to the specified range plus a fixed string > to > >> assure that the data is evenly distributed on each task. > >> > >> > 2. The resources for each task. > >> > >> The resources for each task used the default value. > >> > >> > 3. The intended configuration for the jobs. > >> > >> The parallelism of test Job will be adjusted according to resource > >> conditions to fill the cluster as much as possible. Other configurations > >> are not supported at this time. > >> > >> > 4. What exact source and sink it would use. > >> > >> In order to reduce the dependence on the external, the source data is > >> generated randomly, the sink only supports hdfs or no sinks. > >> > >> we will add this details to the flip laterly. > >> > >> > >> > 在 2019年11月18日，下午7:59，Becket Qin <becket....@gmail.com> 写道： > >> > > >> > +1 (binding) on having the test suite. > >> > > >> > BTW, it would be good to have a few more details about the performance > >> > tests. For example: > >> > 1. How do the testing records look like? The size and key > distributions. > >> > 2. The resources for each task. > >> > 3. The intended configuration for the jobs. > >> > 4. What exact source and sink it would use. > >> > > >> > Thanks, > >> > > >> > Jiangjie (Becket) Qin > >> > > >> > On Mon, Nov 18, 2019 at 7:25 PM Zhijiang <wangzhijiang...@aliyun.com > >> .invalid> > >> > wrote: > >> > > >> >> +1 (binding)! > >> >> > >> >> It is a good thing to enhance our testing work. > >> >> > >> >> Best, > >> >> Zhijiang > >> >> > >> >> > >> >> ------------------------------------------------------------------ > >> >> From:Hequn Cheng <chenghe...@gmail.com> > >> >> Send Time:2019 Nov. 18 (Mon.) 18:22 > >> >> To:dev <dev@flink.apache.org> > >> >> Subject:Re: [VOTE] FLIP-83: Flink End-to-end Performance Testing > >> Framework > >> >> > >> >> +1 (binding)! > >> >> I think this would be very helpful to detect regression problems. > >> >> > >> >> Best, Hequn > >> >> > >> >> On Mon, Nov 18, 2019 at 4:28 PM vino yang <yanghua1...@gmail.com> > >> wrote: > >> >> > >> >>> +1 (non-binding) > >> >>> > >> >>> Best, > >> >>> Vino > >> >>> > >> >>> jincheng sun <sunjincheng...@gmail.com> 于2019年11月18日周一 下午2:31写道： > >> >>> > >> >>>> +1 (binding) > >> >>>> > >> >>>> OpenInx <open...@gmail.com> 于2019年11月18日周一 下午12:09写道： > >> >>>> > >> >>>>> +1 (non-binding) > >> >>>>> > >> >>>>> On Mon, Nov 18, 2019 at 11:54 AM aihua li <liaihua1...@gmail.com> > >> >>> wrote: > >> >>>>> > >> >>>>>> +1 (non-binding) > >> >>>>>> > >> >>>>>> Thanks Yu Li for driving on this. > >> >>>>>> > >> >>>>>>> 在 2019年11月15日，下午8:10，Yu Li <car...@gmail.com> 写道： > >> >>>>>>> > >> >>>>>>> Hi All, > >> >>>>>>> > >> >>>>>>> I would like to start the vote for FLIP-83 [1] which is > discussed > >> >>> and > >> >>>>>>> reached consensus in the discussion thread [2]. > >> >>>>>>> > >> >>>>>>> The vote will be open for at least 72 hours (excluding weekend). > >> >>> I'll > >> >>>>> try > >> >>>>>>> to close it by 2019-11-20 21:00 CST, unless there is an > objection > >> >>> or > >> >>>>> not > >> >>>>>>> enough votes. > >> >>>>>>> > >> >>>>>>> [1] > >> >>>>>>> > >> >>>>>> > >> >>>>> > >> >>>> > >> >>> > >> >> > >> > https://cwiki.apache.org/confluence/display/FLINK/FLIP-83%3A+Flink+End-to-end+Performance+Testing+Framework > >> >>>>>>> [2] https://s.apache.org/7fqrz > >> >>>>>>> > >> >>>>>>> Best Regards, > >> >>>>>>> Yu > >> >>>>>> > >> >>>>>> > >> >>>>> > >> >>>> > >> >>> > >> >> > >> >> > >> > >> >
