[
https://issues.apache.org/jira/browse/KAFKA-13008?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17379462#comment-17379462
]
A. Sophie Blee-Goldman commented on KAFKA-13008:
------------------------------------------------
{quote}Re-reading KIP-227, it seems like there should be a way for the client
to add re-acquired partitions like this to the incremental fetch request so
that it can reinitialize its metadata cache. In other words, it seems like
getting a partition assigned that you haven't owned for a while is effectively
the same case as getting a partition that you've never owned, and there does
seem to be a mechanism for the latter.
{quote}
Thanks John, that is exactly what I was trying to suggest above, but I may have
mungled it with my lack of understanding of the incremental fetch design. Given
how long this bug went unnoticed and the in-depth investigation it took to
uncover the bug (again, nicely done [~showuon]), it seems like any user of the
plain consumer client in addition to Streams could be easily tripped up by
this. And just personally, I had to read the analysis twice to really
understand what was going on, since the behavior was/is so unintuitive to me.
> Stream will stop processing data for a long time while waiting for the
> partition lag
> ------------------------------------------------------------------------------------
>
> Key: KAFKA-13008
> URL: https://issues.apache.org/jira/browse/KAFKA-13008
> Project: Kafka
> Issue Type: Bug
> Affects Versions: 3.0.0
> Reporter: Luke Chen
> Priority: Blocker
> Fix For: 3.0.0
>
> Attachments: image-2021-07-07-11-19-55-630.png
>
>
> In KIP-695, we improved the task idling mechanism by checking partition lag.
> It's a good improvement for timestamp sync. But I found it will cause the
> stream stop processing the data for a long time while waiting for the
> partition metadata.
>
> I've been investigating this case for a while, and figuring out the issue
> will happen in below situation (or similar situation):
> # start 2 streams (each with 1 thread) to consume from a topicA (with 3
> partitions: A-0, A-1, A-2)
> # After 2 streams started, the partitions assignment are: (I skipped some
> other processing related partitions for simplicity)
> stream1-thread1: A-0, A-1
> stream2-thread1: A-2
> # start processing some data, assume now, the position and high watermark is:
> A-0: offset: 2, highWM: 2
> A-1: offset: 2, highWM: 2
> A-2: offset: 2, highWM: 2
> # Now, stream3 joined, so trigger rebalance with this assignment:
> stream1-thread1: A-0
> stream2-thread1: A-2
> stream3-thread1: A-1
> # Suddenly, stream3 left, so now, rebalance again, with the step 2
> assignment:
> stream1-thread1: A-0, *A-1*
> stream2-thread1: A-2
> (note: after initialization, the position of A-1 will be: position: null,
> highWM: null)
> # Now, note that, the partition A-1 used to get assigned to stream1-thread1,
> and now, it's back. And also, assume the partition A-1 has slow input (ex: 1
> record per 30 mins), and partition A-0 has fast input (ex: 10K records /
> sec). So, now, the stream1-thread1 won't process any data until we got input
> from partition A-1 (even if partition A-0 is buffered a lot, and we have
> `{{max.task.idle.ms}}` set to 0).
>
> The reason why the stream1-thread1 won't process any data is because we can't
> get the lag of partition A-1. And why we can't get the lag? It's because
> # In KIP-695, we use consumer's cache to get the partition lag, to avoid
> remote call
> # The lag for a partition will be cleared if the assignment in this round
> doesn't have this partition. check
> [here|https://github.com/apache/kafka/blob/trunk/clients/src/main/java/org/apache/kafka/clients/consumer/internals/SubscriptionState.java#L272].
> So, in the above example, the metadata cache for partition A-1 will be
> cleared in step 4, and re-initialized (to null) in step 5
> # In KIP-227, we introduced a fetch session to have incremental fetch
> request/response. That is, if the session existed, the client(consumer) will
> get the update only when the fetched partition have update (ex: new data).
> So, in the above case, the partition A-1 has slow input (ex: 1 record per 30
> mins), it won't have update until next 30 mins, or wait for the fetch session
> become inactive for (default) 2 mins to be evicted. Either case, the metadata
> won't be updated for a while.
>
> In KIP-695, if we don't get the partition lag, we can't determine the
> partition data status to do timestamp sync, so we'll keep waiting and not
> processing any data. That's why this issue will happen.
>
> *Proposed solution:*
> # If we don't get the current lag for a partition, or the current lag > 0,
> we start to wait for max.task.idle.ms, and reset the deadline when we get the
> partition lag, like what we did in previous KIP-353
> # Introduce a waiting time config when no partition lag, or partition lag
> keeps > 0 (need KIP)
> [~vvcephei] [~guozhang] , any suggestions?
>
> cc [~ableegoldman] [~mjsax] , this is the root cause that in
> [https://github.com/apache/kafka/pull/10736,] we discussed and thought
> there's a data lose situation. FYI.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
