Hi, I have a topic with 50 partitions. This is a development setup so we have 2 days log retention and 1 week offset retention. We had a situation where the consumer service was down for 6 days while the producer was still producing records to the topic. The consumer is configure to "latest" for auto.offset.reset. When the consumer came up after 6 days, it resumed consumption of data from 28 of the 50 partitions from where it was left off and for other partitions it went to latest. The exported data from s3 connector task (which was always running) revealed the there was data for all partitions during those 6 days. It looked like the log was retained for those 28 partitions although it was set for 2 days while the log for other partitions was retained for last tow days. Is this expected behavior under any settings or scenarios. Was wondering how to troubleshoot it further. Any insights would help.
Thanks, -Manu
