[
https://issues.apache.org/jira/browse/KAFKA-2801?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14999595#comment-14999595
]
ASF GitHub Bot commented on KAFKA-2801:
---------------------------------------
GitHub user rajinisivaram opened a pull request:
https://github.com/apache/kafka/pull/493
KAFKA-2801: Process any remaining data in SSL network read buffer after
handshake
Process any remaining data in the network read buffer in
`SslTransportLayer` when `read()` is invoked. On handshake completion, there
could be application data ready to be processed that was read into
`netReadBuffer` during handshake processing. `read()` is already invoked from
`Selector` after handshake completion, but data already read into the
`netReadBuffer` was not being processed. This PR adds a check for remaining
data and continues with processing data if data is available.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/rajinisivaram/kafka KAFKA-2801
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/kafka/pull/493.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #493
----
commit 081cdc953c04e0fa12c7c8633c9f2d5e1db8638f
Author: Rajini Sivaram <rajinisiva...@googlemail.com>
Date: 2015-11-10T23:19:37Z
KAFKA-2801: Process any remaining data in network read buffer during SSL
read
----
> Data read from network not processed by SSL transport layer
> -----------------------------------------------------------
>
> Key: KAFKA-2801
> URL: https://issues.apache.org/jira/browse/KAFKA-2801
> Project: Kafka
> Issue Type: Bug
> Components: network
> Affects Versions: 0.9.0.0
> Reporter: Rajini Sivaram
> Assignee: Rajini Sivaram
> Priority: Blocker
> Fix For: 0.9.0.0
>
>
> We have been seeing intermittent failures in our performance tests when
> producer times out while waiting for metadata response, when running with
> SSL. Digging deeper into this failure, this is a result of data that was read
> into _SslTransporLayer.netReadBuffer_ during handshake processing, but never
> processed later unless more data arrives on the network. In the case of the
> producer, no more data is sent until the metadata response is received and
> hence metadata request is never processed, leading to timeouts in the
> producer.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
