[ 
https://issues.apache.org/jira/browse/STORM-1632?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15209337#comment-15209337
 ] 

ASF GitHub Bot commented on STORM-1632:
---------------------------------------

Github user HeartSaVioR commented on the pull request:

    https://github.com/apache/storm/pull/1217#issuecomment-200571684
  
    @roshannaik 
    I'm sorry I feel I picked the wrong word `micro optimization` to confuse 
you what I mean. `local optimization` seems clearer.
    
    Btw, I guess @ptgoetz got me.
    Yes, I agree we need micro-benchmark to clear out variables, but I think it 
has to be re-evaluated with normal benchmark to reason how it affects in 
relatively normal situation, especially if it has to touch functionalities.
    If it doesn't touch functionality I would say "Awesome work!" even though 
under 5% of performance gain on local optimization.
    (Why STORM-1526 and STORM-1539 didn't need to re-evaluate with normal 
benchmark is that it didn't affect any functionalities.)
    
    And I guess this overhead (0.006720889 ms = 6720.889 ns per each tuple 
spend in send_to_eventlogger  as @arunmahadevan posted) is relatively very 
small than what Storm has to do for process tuple - enqueue and dequeue, 
finding task id to send, serde, transfer - which we may find spots to improve.
    
    Though I agree that's inside of critical path so we may want to find the 
alternative way with not touching functionality. 
    If you really want to disable it by default, it would be better to post 
mail to dev mailing list to build consensus first.


> Disable event logging by default
> --------------------------------
>
>                 Key: STORM-1632
>                 URL: https://issues.apache.org/jira/browse/STORM-1632
>             Project: Apache Storm
>          Issue Type: Bug
>          Components: storm-core
>            Reporter: Roshan Naik
>            Assignee: Roshan Naik
>            Priority: Blocker
>             Fix For: 1.0.0
>
>
> EventLogging has performance penalty. For a simple speed of light topology  
> with a single instances of a spout and a bolt, disabling event logging 
> delivers a 7% to 9% perf improvement (with acker count =1)
> Event logging can be enabled when there is need to do debug, but turned off 
> by default.
> **Update:** with acker=0  the observed impact was much higher... **25%** 
> faster when event loggers = 0 



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

Reply via email to