[
https://issues.apache.org/jira/browse/KAFKA-3442?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15206867#comment-15206867
]
Jun Rao commented on KAFKA-3442:
--------------------------------
[~becket_qin], it's probably better if we keep supporting the current partial
message behavior. If the non-java clients follow the same logic as the java
client for dealing with partial messages, I was thinking that if the broker
detects this case, it can just send a constant ByteBufferMessageSet that
contains an arbitrary partial message.
[~dana.powers], it's probably a good idea for the broker to detect the partial
message case and send an error in the response directly instead of having the
client to guess this from the payload. I think this is possible since the
broker already does a small scan to find the first message to return. We can
just further check the size of the first message. It's too late to do this in
0.10.0 though since this may require bumping the version of FetchRequest.
[~ijuma], do you want to file a separate jira to track this?
> FetchResponse size exceeds max.partition.fetch.bytes
> ----------------------------------------------------
>
> Key: KAFKA-3442
> URL: https://issues.apache.org/jira/browse/KAFKA-3442
> Project: Kafka
> Issue Type: Bug
> Affects Versions: 0.10.0.0
> Reporter: Dana Powers
> Assignee: Jiangjie Qin
> Priority: Blocker
> Fix For: 0.10.0.0
>
>
> Produce 10000 byte message to topic foobar
> Fetch foobar w/ max.partition.fetch.bytes=1024
> Test expects to receive a truncated message (~1024 bytes). 0.8 and 0.9 pass
> this test, but 0.10 FetchResponse has full message, exceeding the max
> specified in the FetchRequest.
> I tested with v0 and v1 apis, both fail. Have not tested w/ v2
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
