luoyuxia commented on PR #2185: URL: https://github.com/apache/fluss/pull/2185#issuecomment-3754632952
> > > Thanks @luoyuxia for the work, I left some minor comment. > > > But I also have a question for current design, Could we address large splits case via current design? The data freshness would be worse for large log split from my understanding. > > > > > > Hi, thanks for reviewing my pr. Yes, large log split will harm data fressnes. But What do you mean `address large splits case`? In the current design, if we have a large split, and aussming the max tiering duration is 10 min, after 10 min, the large split will be forced to commit. > > I mean for the large log table, the data freshness goal(i.e. 5minute) would be worse to 10 minutes, 15 minutes after some scheduled rounds, maybe in that case we need monitor these tables and scale up tiering service resources. Yes, your are right. But IIUC, the data freshness goal is a best effort when tiering resource is enough, but when tiering resource is not enough, it's normal that we can't reach to the goal. But a round of tiering of a table reach the max duraiton, it means that the resource is not enough, we need to scale up tiering service resources.. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
