No worries. If you could share how far you got and what the next steps are and if you still think its the right direction, it ll help the person picking this up :)
On Fri, Dec 11, 2020 at 3:52 AM vino yang <[email protected]> wrote: > Hi, > > Sorry. > Due to busy work, I interrupted this verification work. I will let > @wangxianghu continue to follow up this work. > > Best, > Vino > > Vinoth Chandar <[email protected]> 于2020年12月11日周五 下午3:25写道： > > > Thanks for kicking this off, Gary! > > > > Thanks to Raymond, the CI tests are leaner and more parellelized now. > But I > > do see that we hit travis queuing a lot. > > Is that what you are targeting primarily? > > > > We were exploring a move to Azure CI Pipelines to get more compute power > > allocated at some point. > > @vinoyang , IIRC, you were driving this? Could you please share your > > experience? > > > > On Thu, Dec 10, 2020 at 9:53 PM Gary Li <[email protected]> wrote: > > > > > Hi all, > > > > > > I am seeing we have a boost on PRs recently. That's great news that > more > > > developers joined the Hudi community. > > > > > > On the other hand, the Travis build backlog was growing as well. As we > > > grow, we need to scale the CI/CD pipeline soon. So I'd like to start a > > > discussion here. > > > > > > Do we have any ongoing work on this topic? Anyone from larger open > source > > > projects like Spark/Flink e.t.c can share experience about how they > > handle > > > this problem? > > > > > > Any thoughts are appreciated. > > > > > > Best, > > > > > > Gary Li > > > > > >
