> 2. "This is a breaking change because it alters where these two configs are applied. The change can only be introduced in the next major release (i.e., 5.0)." Does that mean we don't do anything before v5.0? Maybe we can update the doc first to mention this inconsistency?
Yes, documentation is a good point. Also, it would be better to print some warning message if broker has non-default partitions and replicas in the static configs Luke Chen <[email protected]> 於 2025年9月25日 週四 下午3:01寫道： > Hi Kuan-Po, > > Thanks for fixing this inconsistency! > > Questions: > 1. The `num.partitions` currently still exists in our example > broker.properties file. We should update it, too. > > 2. "This is a breaking change because it alters where these two configs are > applied. The change can only be introduced in the next major release (i.e., > 5.0)." > Does that mean we don't do anything before v5.0? Maybe we can update the > doc first to mention this inconsistency? > > 3. What about the topic configuration overridden in broker and controller > properties file? > For example: > Broker sets "log.retention.ms=100" > Controller sets "log.retention.ms=1000" > A topic "t" created without setting the "retention.ms". > In this case, the "retention.ms" for topic "t" is 100 or 1000? > Does the inconsistency happen in all topic configs? > > > Thank you. > Luke > > > On Thu, Sep 25, 2025 at 1:40 AM Chia-Ping Tsai <[email protected]> wrote: > > > hi Kuan-Po > > > > thanks for this KIP. > > > > chia_00: Is it possible to add warning messages to broker to highlight > this > > change? > > > > chia_01: Should broker set `-1` explicitly? > > > > Best, > > Chia-Ping > > > > Kuan-Po Tseng <[email protected]> 於 2025年9月25日 週四 上午12:20寫道： > > > > > Hello everyone, > > > > > > I’d like to bring up a small change aimed at aligning the behavior of > > > num.partitions and default.replication.factor during topic creation. > > > > > > You can find the KIP here: > https://cwiki.apache.org/confluence/x/WIrHFg. > > > > > > Thanks, > > > Kuan-Po Tseng > > > > > >
