-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/20050/#review39825
-----------------------------------------------------------



clients/src/main/java/org/apache/kafka/clients/producer/internals/Sender.java
<https://reviews.apache.org/r/20050/#comment72489>

    Perhaps these can all be wrapped in the same if statement in line 815 to 
save the unnecessary overhead?



clients/src/main/java/org/apache/kafka/clients/producer/internals/Sender.java
<https://reviews.apache.org/r/20050/#comment72490>

    Ditto as the above.



clients/src/main/java/org/apache/kafka/clients/producer/internals/Sender.java
<https://reviews.apache.org/r/20050/#comment72494>

    I am wondering if this is the right place to do the registration. Since 
this is only called when we have producer request, if the first request sent to 
a node is not a produce request, we won't have pre-created the node level 
sensor. I think this can happen during a broker failure when the first request 
to a node could be a metadata request.
    
    So, we probably have to register this in Selector. This may also solve your 
other problem that a metric is not registered in unit test.


- Jun Rao


On April 4, 2014, 9:33 p.m., Guozhang Wang wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/20050/
> -----------------------------------------------------------
> 
> (Updated April 4, 2014, 9:33 p.m.)
> 
> 
> Review request for kafka.
> 
> 
> Bugs: KAFKA-1359
>     https://issues.apache.org/jira/browse/KAFKA-1359
> 
> 
> Repository: kafka
> 
> 
> Description
> -------
> 
> KAFKA-1359.v1
> 
> 
> Diffs
> -----
> 
>   
> clients/src/main/java/org/apache/kafka/clients/producer/internals/Sender.java 
> 855ae84f14aa91653b3fa855c2af40560323f42a 
>   clients/src/main/java/org/apache/kafka/common/network/Selector.java 
> 558f8b47638b354f9c1a30be5d45dc7b61131bea 
> 
> Diff: https://reviews.apache.org/r/20050/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Guozhang Wang
> 
>

Reply via email to