[
https://issues.apache.org/jira/browse/KAFKA-202?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13183482#comment-13183482
]
Neha Narkhede edited comment on KAFKA-202 at 1/10/12 7:07 PM:
--------------------------------------------------------------
I tried to apply the v5 patch and couldn't find RequestChannel.scala in there.
Would you mind uploading a patch with that included ?
Also, if the kafka cluster is experiencing either network or IO bottleneck, it
slows processing down and backs up the producer queue causing
QueueFullException. To detect this, seems like it will be helpful to expose the
number of queued requests in the SocketServerStats ? This could either be a
separate JIRA or part of this one, your call.
was (Author: nehanarkhede):
I tried to apply the v5 patch and couldn't find RequestChannel.scala in
there. Would you mind uploading a patch with that included ?
Also, if the kafka cluster is experiencing either network or IO bottleck, it
slows processing down and backs up the producer queue causing
QueueFullException. To detect this, seems like it will be helpful to expose the
number of queued requests in the SocketServerStats ? This could either be a
separate JIRA or part of this one, your call.
> Make the request processing in kafka asynchonous
> ------------------------------------------------
>
> Key: KAFKA-202
> URL: https://issues.apache.org/jira/browse/KAFKA-202
> Project: Kafka
> Issue Type: New Feature
> Reporter: Jay Kreps
> Assignee: Jay Kreps
> Attachments: KAFKA-202-v2.patch, KAFKA-202-v3.patch,
> KAFKA-202-v4.patch, KAFKA-202-v5.patch,
> KAFKA-48-socket-server-refactor-draft.patch
>
>
> We need to handle long-lived requests to support replication. To make this
> work we need to make the processing mechanism asynchronous from the network
> threads.
> To accomplish this we will retain the existing pool of network threads but
> add a new pool of request handling threads. These will do all the disk I/O.
> There will be a queue mechanism to transfer requests to and from this
> secondary pool.
--
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
