kafka-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Neha Narkhede (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (KAFKA-449) Leader election test
Date Tue, 11 Sep 2012 16:39:08 GMT

    [ https://issues.apache.org/jira/browse/KAFKA-449?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13453163#comment-13453163

Neha Narkhede commented on KAFKA-449:

Overall, looks good. Here are a few minor suggestions -

1. Let's try to reduce the amount of sleep statements in the test. It runs pretty slowly right
2. I see that the dashboards path point to some Hudson specific directory. This might break
the metrics dashboards on local tests. Let's fix that
3. Let's reduce the amount of INFO logging. For example, we don't need to the unix timestamp
on the leader or the leader election latency while running the test. You can summarize that
at the end of the test run. Same goes for the producer "status of stopBackgroundProducer"
> Leader election test
> --------------------
>                 Key: KAFKA-449
>                 URL: https://issues.apache.org/jira/browse/KAFKA-449
>             Project: Kafka
>          Issue Type: Sub-task
>    Affects Versions: 0.8
>            Reporter: Neha Narkhede
>            Assignee: John Fung
>              Labels: replication-testing
>         Attachments: kafka-449-v1.patch
>   Original Estimate: 24h
>  Remaining Estimate: 24h
> The goal is to ensure that leader election completely successfully each time and also
record the leader election latency. Setup multiple partitions with a particular replication
factor. The test should take in the number of iterations. In each iteration, it bounces one
server. That server will be the leader for a certain set of partitions.  At the end of the
test, report the leader election latency along with message replication verification. The
test consists of several runs, each with a particular replication factor (between 1 and 6)

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

View raw message