kafka-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jay Kreps (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (KAFKA-2486) New consumer performance
Date Fri, 28 Aug 2015 17:02:46 GMT

    [ https://issues.apache.org/jira/browse/KAFKA-2486?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14720237#comment-14720237
] 

Jay Kreps commented on KAFKA-2486:
----------------------------------

[~guozhang] I interpret you to be saying that even if there is data, sometimes fetching immediately
isn't ideal for throughput because of the reduction in batching? I think that is right but
that is controlled by the min bytes setting in the fetch request, right? In other words I
think we already address that. Not sure what the ideal consumer perf settings are. I think
for that perf test it shoudn't matter too much what the max wait is because there is always
data to fetch so it never really get's used.

I agree that right now letting the consumer set the max wait to 0 does allow the possibility
of abuse but I think the fix for that is probably an eventual server-side quota.

> New consumer performance
> ------------------------
>
>                 Key: KAFKA-2486
>                 URL: https://issues.apache.org/jira/browse/KAFKA-2486
>             Project: Kafka
>          Issue Type: Sub-task
>          Components: consumer
>            Reporter: Ewen Cheslack-Postava
>            Assignee: Jason Gustafson
>             Fix For: 0.8.3
>
>
> The new consumer was previously reaching getting good performance. However, a recent
report on the mailing list indicates it's dropped significantly. After evaluation, even with
a local broker it seems to only be reaching a 2-10MB/s, compared to 600+MB/s previously. Before
release, we should get the performance back on par.
> Some details about where the regression occurred from the mailing list http://mail-archives.apache.org/mod_mbox/kafka-dev/201508.mbox/%3CCAAdKFaE8bPSeWZf%2BF9RuA-xZazRpBrZG6vo454QLVHBAk_VOJg%40mail.gmail.com%3E
:
> bq. At 49026f11781181c38e9d5edb634be9d27245c961 (May 14th), we went from good performance
-> an error due to broker apparently not accepting the partition assignment strategy. Since
this commit seems to add heartbeats and the server side code for partition assignment strategies,
I assume we were missing something on the client side and by filling in the server side, things
stopped working.
> bq. On either 84636272422b6379d57d4c5ef68b156edc1c67f8 or a5b11886df8c7aad0548efd2c7c3dbc579232f03
(July 17th), I am able to run the perf test again, but it's slow -- ~10MB/s for me vs the
2MB/s Jay was seeing, but that's still far less than the 600MB/s I saw on the earlier commits.
> Ideally we would also at least have a system test in place for the new consumer, even
if regressions weren't automatically detected. It would at least allow for manually checking
for regressions. This should not be difficult since there are already old consumer performance
tests.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message