[ https://issues.apache.org/jira/browse/KAFKA-5002?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15953717#comment-15953717 ]
Matthias J. Sax commented on KAFKA-5002: ---------------------------------------- "if i start consumer which consumes data from topic t1 and partition p1 then the stream logic doesn't consider p1 for processing" -> starting a consumer (with different {{group.id}}) should not conflict with Kafka Streams -- it would be invalid (and should fail with an exception) if you start a consumer with the same {{group.id}} as the Kafka Streams app (note, Kafka Streams uses it's {{application.id}} as {{group.id}} internally). Just want to verify, I understand correctly what you are reporting. > Stream does't seem to consider partitions for processing which are being > consumed > --------------------------------------------------------------------------------- > > Key: KAFKA-5002 > URL: https://issues.apache.org/jira/browse/KAFKA-5002 > Project: Kafka > Issue Type: Bug > Components: streams > Affects Versions: 0.10.2.0 > Environment: Windows 8.1 > Reporter: Mustak > Labels: patch > > Kafka streams doesn't seems to consider particular partition for processing > if that partition is being consumed by some consumer. For example if I've two > topics t1 and t2 with two partitions p1 and p2 and there is a stream process > is running with consumes data from these topics and produce output to topic > t3 which has two partitions. If run this kind of topology it works but if i > start consumer which consumes data from topic t1 and partition p1 then the > stream logic doesn't consider p1 for processing and stream doesn't provide > any output related to that partition. I think stream logic should consider > partitions which are being consumed. -- This message was sent by Atlassian JIRA (v6.3.15#6346)