[
https://issues.apache.org/jira/browse/KAFKA-8315?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16841213#comment-16841213
]
Andrew commented on KAFKA-8315:
-------------------------------
[~vvcephei] [~ableegoldman] Using the integration test I think I now understand
what is going on.
The key bit of code is here :
[https://github.com/the4thamigo-uk/kafka/blob/trunk/streams/src/main/java/org/apache/kafka/streams/processor/internals/StreamTask.java#L332]
What appears to be a happening is this :
1) Since the topics are already full of data, the left topic has sufficient
data (1000 records) in order to trigger leaving this loop
[https://github.com/the4thamigo-uk/kafka/blob/debugging/clients/src/main/java/org/apache/kafka/clients/consumer/internals/Fetcher.java#L485.]
So, no right records are fetched.
2) All the fetched left stream records are added to PartitionGroup, and
PartitionGroup.allBuffered = false, since the right stream RecordQueue is still
empty
3) The code drops into here
[https://github.com/the4thamigo-uk/kafka/blob/debugging/streams/src/main/java/org/apache/kafka/streams/processor/internals/StreamTask.java#L330,]
since maxTaskIdleMs == 0 !
4) The 1000 left stream records are processed (thereby almost immediately
expiring their join windows !
5) The right stream records are fetched and processed, but there are no left
stream join windows to join with until the latest records in the left stream
for which the windows have not expired.
And the workaround/fix, is a change of configuration setting :
[https://github.com/the4thamigo-uk/kafka/commit/189aa764aef06643a8a3c30b2aee3c4a29b82ae6]
Perhaps this value should default to non-zero to enable historical joins by
default?
> Historical join issues
> ----------------------
>
> Key: KAFKA-8315
> URL: https://issues.apache.org/jira/browse/KAFKA-8315
> Project: Kafka
> Issue Type: Bug
> Components: streams
> Reporter: Andrew
> Assignee: John Roesler
> Priority: Major
> Attachments: code.java
>
>
> The problem we are experiencing is that we cannot reliably perform simple
> joins over pre-populated kafka topics. This seems more apparent where one
> topic has records at less frequent record timestamp intervals that the other.
> An example of the issue is provided in this repository :
> [https://github.com/the4thamigo-uk/join-example]
>
> The only way to increase the period of historically joined records is to
> increase the grace period for the join windows, and this has repercussions
> when you extend it to a large period e.g. 2 years of minute-by-minute records.
> Related slack conversations :
> [https://confluentcommunity.slack.com/archives/C48AHTCUQ/p1556799561287300]
> [https://confluentcommunity.slack.com/archives/C48AHTCUQ/p1557733979453900]
>
> Research on this issue has gone through a few phases :
> 1) This issue was initially thought to be due to the inability to set the
> retention period for a join window via {{Materialized: i.e.}}
> The documentation says to use `Materialized` not `JoinWindows.until()`
> ([https://kafka.apache.org/22/javadoc/org/apache/kafka/streams/kstream/JoinWindows.html#until-long-]),
> but there is no where to pass a `Materialized` instance to the join
> operation, only to the group operation is supported it seems.
> This was considered to be a problem with the documentation not with the API
> and is addressed in [https://github.com/apache/kafka/pull/6664]
> 2) We then found an apparent issue in the code which would affect the
> partition that is selected to deliver the next record to the join. This would
> only be a problem for data that is out-of-order, and join-example uses data
> that is in order of timestamp in both topics. So this fix is thought not to
> affect join-example.
> This was considered to be an issue and is being addressed in
> [https://github.com/apache/kafka/pull/6719]
> 3) Further investigation using a crafted unit test seems to show that the
> partition-selection and ordering (PartitionGroup/RecordQueue) seems to work ok
> [https://github.com/the4thamigo-uk/kafka/commit/5121851491f2fd0471d8f3c49940175e23a26f1b]
> 4) the current assumption is that the issue is rooted in the way records are
> consumed from the topics :
> We have tried to set various options to suppress reads form the source topics
> but it doesnt seem to make any difference :
> [https://github.com/the4thamigo-uk/join-example/commit/c674977fd0fdc689152695065d9277abea6bef63]
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
