Hi, Dong, Thanks for the KIP. Looks good overall. Just a few more comments.
1."Note that the way broker handles PurgeRequest is similar to how it handles ProduceRequest with ack = -1 and isr=all_replicas". It seems that the implementation is a bit different. In this KIP, we wait for all replicas. But in producer, acks=all means waiting for all in-sync replicas. 2. Could you describe the behavior when the specified offsetToPurge is (a) smaller than lowWatermark, (b) larger than highWatermark, but smaller than log end offset, (c) larger than log end offset? 3. In the ListOffsetRequest, will lowWatermark be returned when the smallest_offset option is used? Thanks, Jun On Wed, Jan 11, 2017 at 1:01 PM, Dong Lin <lindon...@gmail.com> wrote: > Hi all, > > It seems that there is no further concern with the KIP-107. At this point > we would like to start the voting process. The KIP can be found at > https://cwiki.apache.org/confluence/display/KAFKA/KIP-107 > %3A+Add+purgeDataBefore%28%29+API+in+AdminClient. > > Thanks, > Dong >