[
https://issues.apache.org/jira/browse/KAFKA-2063?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15403700#comment-15403700
]
Ismael Juma commented on KAFKA-2063:
------------------------------------
[~nepal], I believe the answer is yes to 1) and 2). A few points (you may be
aware of them already, but just want to be clear):
1. A change like this needs to follow the KIP process. Would you be willing to
submit a KIP for discussion?
2. Any change we make must take compatibility into account. So, even though we
want the new version of the fetch request not to have the partition level
limit, the broker will still have to support the partition level limit for the
older versions of the fetch request. This is true for both when it receives the
fetch request in KafkaApis as well as when it sends the fetch request in
ReplicaFetcherThread. In the latter case, we will have to use the
inter.broker.protocol.version property to figure out which version of the fetch
request to send.
3. It makes sense to have a new property for the new request level limit. I
think we need it on the broker and client as well, right? Also, if we deprecate
the partition-level properties, we need to think through the consequences for
users who have custom values for these properties and if we can mitigate these
potential issues. This is always important, but it's of particular importance
if we want to include this in a minor release (e.g. 0.10.1.0).
> Bound fetch response size
> -------------------------
>
> Key: KAFKA-2063
> URL: https://issues.apache.org/jira/browse/KAFKA-2063
> Project: Kafka
> Issue Type: Improvement
> Reporter: Jay Kreps
>
> Currently the only bound on the fetch response size is
> max.partition.fetch.bytes * num_partitions. There are two problems:
> 1. First this bound is often large. You may chose
> max.partition.fetch.bytes=1MB to enable messages of up to 1MB. However if you
> also need to consume 1k partitions this means you may receive a 1GB response
> in the worst case!
> 2. The actual memory usage is unpredictable. Partition assignment changes,
> and you only actually get the full fetch amount when you are behind and there
> is a full chunk of data ready. This means an application that seems to work
> fine will suddenly OOM when partitions shift or when the application falls
> behind.
> We need to decouple the fetch response size from the number of partitions.
> The proposal for doing this would be to add a new field to the fetch request,
> max_bytes which would control the maximum data bytes we would include in the
> response.
> The implementation on the server side would grab data from each partition in
> the fetch request until it hit this limit, then send back just the data for
> the partitions that fit in the response. The implementation would need to
> start from a random position in the list of topics included in the fetch
> request to ensure that in a case of backlog we fairly balance between
> partitions (to avoid first giving just the first partition until that is
> exhausted, then the next partition, etc).
> This setting will make the max.partition.fetch.bytes field in the fetch
> request much less useful and we should discuss just getting rid of it.
> I believe this also solves the same thing we were trying to address in
> KAFKA-598. The max_bytes setting now becomes the new limit that would need to
> be compared to max_message size. This can be much larger--e.g. setting a 50MB
> max_bytes setting would be okay, whereas now if you set 50MB you may need to
> allocate 50MB*num_partitions.
> This will require evolving the fetch request protocol version to add the new
> field and we should do a KIP for it.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
