[ https://issues.apache.org/jira/browse/KAFKA-10136?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17138169#comment-17138169 ]
Chia-Ping Tsai commented on KAFKA-10136: ---------------------------------------- {quote} If the option [numberMessages] is "total messages to consume for all threads" {quote} from [https://github.com/apache/kafka/blob/0.9.0/core/src/main/scala/kafka/tools/ConsumerPerformance.scala#L50] It was "total messages to consume for all threads" :) {quote} then an atomic variable is needed to synchronize how many messages have been consumed by all threads and poor performance will be achieved. {quote} You can use LongAdder if you do care the performance of count but adding something to AtomicLong is faster than fetching data from remote servers I'd say :) > Make option threads of ConsumerPerformance work > ----------------------------------------------- > > Key: KAFKA-10136 > URL: https://issues.apache.org/jira/browse/KAFKA-10136 > Project: Kafka > Issue Type: Bug > Reporter: jiamei xie > Assignee: jiamei xie > Priority: Major > > Make option threads of ConsumerPerformance work -- This message was sent by Atlassian Jira (v8.3.4#803005)