Hello jun, jr2, If the same config appears multiple times within a single request, the later value overwrites the earlier value. I have updated this behavior in the KIP.
Best Regards, Jiunn-Yang > Jun Rao via dev <[email protected]> 於 2026年2月4日 凌晨2:30 寫道： > > Hi, Jiunn-Yang, > > Thanks for the reply. One more comment. > > jr2. Duplicate configuration handling: Controllers allow this behavior > without throwing any exception. > What's the impact of the controller allowing duplicate configuration > entries? > > Jun > > > On Tue, Feb 3, 2026 at 2:57 AM 黃竣陽 <[email protected]> wrote: > >> Hello Jun, >> >> jr1. I have updated the KIP >> >> Best Regards, >> Jiunn-Yang >> >>> Jun Rao via dev <[email protected]> 於 2026年2月3日 凌晨2:41 寫道： >>> >>> Hi, Jiunn-Yang, >>> >>> Thanks for the reply. >>> >>> jr1. Could you update the KIP to make the discrepancy and the impact >> clear? >>> >>> Jun >>> >>> On Sat, Jan 31, 2026 at 6:27 PM 黃竣陽 <[email protected]> wrote: >>> >>>> Hello Jun >>>> >>>> jr1. Yes. When using the controller path, AlterConfigOp.OpType.SET >> allows >>>> null values. >>>> As a result, the user’s configuration may be deleted unintentionally. >>>> >>>> Best Regards, >>>> Jiunn-Yang >>>> >>>>> Jun Rao via dev <[email protected]> 於 2026年1月31日 上午8:02 寫道： >>>>> >>>>> Hi, Jiunn-Yang, >>>>> >>>>> Thanks for the KIP. It's important to make the config behavior >> consistent >>>>> between the broker and the controller. >>>>> >>>>> jr1. "Brokers only allow null values with the >> AlterConfigOp.OpType.DELETE >>>>> operation type; otherwise, they throw an InvalidRequestException . >>>>> Controllers allow users to set null values with >>>>> AlterConfigOp.OpType.DELETE without restrictions." >>>>> Do you mean that the controller allow users to set null values for >> types >>>>> other than AlterConfigOp.OpType.DELETE? If that's the case, does that >>>> lead >>>>> to accidentally deleting a config? >>>>> >>>>> Jun >>>>> >>>>> >>>>> >>>>> On Tue, Jan 6, 2026 at 2:02 AM 黃竣陽 <[email protected]> wrote: >>>>> >>>>>> Hi all, >>>>>> >>>>>> I’d like to manually bump this thread. >>>>>> >>>>>> Best Regards, >>>>>> Jiunn-Yang >>>>>> >>>>>>> 黃竣陽 <[email protected]> 於 2025年12月22日 晚上9:25 寫道： >>>>>>> >>>>>>> Hello chia, >>>>>>> >>>>>>> Thanks for your comments. >>>>>>> I have updated the KIP based on your feedback. >>>>>>> >>>>>>> Best Regards, >>>>>>> Jiunn-Yang >>>>>>> >>>>>>>> Chia-Ping Tsai <[email protected]> 於 2025年12月21日 晚上9:40 寫道： >>>>>>>> >>>>>>>> hi Jiunn-Yang >>>>>>>> >>>>>>>> chia_00: please update the discussion link on the KIP page >>>>>>>> >>>>>>>> chia_01: The description of `Invalid dynamic configs` is incomplete. >>>>>> The controller accepts the invalid request initially, but silently >> drops >>>>>> the invalid configs during application. Consequently, previous valid >>>>>> configs are also removed ... >>>>>>>> >>>>>>>> Best, >>>>>>>> Chia-Ping >>>>>>>> >>>>>>>> On 2025/12/15 12:14:52 黃竣陽 wrote: >>>>>>>>> Hello everyone, >>>>>>>>> >>>>>>>>> I would like to start a discussion on KIP-1256 Align broker and >>>>>> controller behavior for handling >>>>>>>>> null config values, duplicate entries, and error codes < >>>>>> https://cwiki.apache.org/confluence/x/J47MFw> >>>>>>>>> >>>>>>>>> This proposal aims to align the behavior of >>>>>> Admin.incrementalAlterConfigs when using bootstrap.servers >>>>>>>>> versus bootstrap.controllers. >>>>>>>>> >>>>>>>>> Best Regards, >>>>>>>>> Jiunn-Yang >>>>>>> >>>>>> >>>>>> >>>> >>>> >> >>
