[
https://issues.apache.org/jira/browse/KAFKA-49?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13222975#comment-13222975
]
Prashanth Menon commented on KAFKA-49:
--------------------------------------
Sounds good to me. Doing some additional work on DefaultEventHandler, I
noticed something off in the retry logic that I'd like to get confirmed.
Consider the case where I've got data destined for more than one broker, say
three.
- Enter handleSerializedData()
- Partioning and collating makes a map with three key/value pairs (broker ->
topic partition data and messages).
- Enter for loop
- Assume the first send works on the first try for broker #1.
- Next iteration, the second send to broker #2 fails on the first try, we
fall into the retry loop and recursive into handleSerializedData with
requiredRetries = 0.
- In handleSerializedData
- This time, the partitioned data will one one key/value pair for the
single broker (broker #2) we're attempting to resend data to.
- Enter for loop
- Attempt to send data to broker #2, the send succeeds
- We exhaust the map entries and the for loop condition.
- We return to the retry loop for retry=1 on broker #2 in the catch block.
- The previous send succeeded on first try and now there's the "return"
statement. This exists the function, but we have one more broker (broker #3)
to send data to.
Does the flow sound about right? I think what needs to happen is to set a flag
and break the retry while loop after a successful retry. Then we check the
flag after the loop and either throw the exception or continue the outer for
loop.
Am I crazy? Am I missing something in my sleep-deprived state here?
> Add acknowledgement to the produce request.
> -------------------------------------------
>
> Key: KAFKA-49
> URL: https://issues.apache.org/jira/browse/KAFKA-49
> Project: Kafka
> Issue Type: Bug
> Reporter: Jun Rao
> Assignee: Prashanth Menon
>
> Currently, the produce request doesn't get acknowledged. We need to have a
> broker send a response to the producer and have the producer wait for the
> response before sending the next request.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
