Thanks, Bruno, I think my feelings are the same as yours. It seems like either call is just a matter of some forward looking statement in the KIP and maybe a warning log if we're leaning toward changing the default in the future.
I'm happy with whatever you prefer. Thanks again, -John On Thu, 2020-12-10 at 16:37 +0100, Bruno Cadonna wrote: > Hi John, > > Thank you for the feedback! > > I am undecided, because while manual init only makes Kafka Streams safer > regarding data loss, it makes first toy apps with Kafka Streams a little > bit more complicated. I am a bit more inclined to manual init only, though. > > Best, > Bruno > > > On 10.12.20 15:20, John Roesler wrote: > > Hi Bruno, > > > > Thanks for the KIP! > > > > This seems like a nice data integrity improvement, and the KIP looks good > > to me. > > > > I’m wondering if we should plan to transition to manual init only in the > > future. I.e. maybe we log a warning, then later on we switch the default > > config to manual, and then ultimately drop the config completely. What do > > you think? > > > > Thanks, > > John > > > > On Thu, Dec 10, 2020, at 04:36, Bruno Cadonna wrote: > > > Hi, > > > > > > I'd like to start the discussion on KIP-698 that proposes an explicit > > > user initialization of broker-side state for Kafka Streams instead of > > > letting Kafka Streams setting up the broker-side state automatically > > > during rebalance. Such an explicit initialization avoids possible data > > > loss issues due to automatic initialization. > > > > > > https://cwiki.apache.org/confluence/x/7CnZCQ > > > > > > Best, > > > Bruno > > >
