Author: Dong Lin <lindong28@gmail.com>
Reviewers: Ewen Cheslack-Postava <me@ewencp.org>, Jason Gustafson <jason@confluent.io>
Closes#2155 from lindong28/KAFKA-4429
The original Javadoc description for `ConsumerRecord` is slightly confusing in that it can be read in a way such that an object is a key value pair received from Kafka, but (only) consists of the metadata associated with the record. This PR makes it clearer that the metadata is _included_ with the record, and moves the comma so that the phrase "topic name and partition number" in the sentence is more closely associated with the phrase "from which the record is being received".
Author: LoneRifle <LoneRifle@users.noreply.github.com>
Reviewers: Ismael Juma <ismael@juma.me.uk>, Ewen Cheslack-Postava <ewen@confluent.io>
Closes#2290 from LoneRifle/patch-1
Author: MURAKAMI Masahiko <fossamagna2@gmail.com>
Reviewers: Jun Rao <junrao@gmail.com>, Ismael Juma <ismael@juma.me.uk>
Closes#2265 from fossamagna/fix-lz4outputstream-close
This was changed in b58b6a1bef and caused the `ReplicaVerificationToolTest.test_replica_lags`
system test to start failing.
I also added a unit test and a couple of other minor clean-ups.
Author: Ismael Juma <ismael@juma.me.uk>
Reviewers: Jason Gustafson <jason@confluent.io>
Closes#2280 from ijuma/kafka-4554-fix-replica-buffer-verify-checksum
In case of file record truncation during write due to improper types usage
(`AtomicInteger` in place of `int`) `IllegalFormatConversionException` would
be thrown instead of `KafkaException`
Author: Kamil Szymanski <kamil.szymanski.dev@gmail.com>
Reviewers: Ismael Juma <ismael@juma.me.uk>
Closes#2275 from kamilszymanski/file_record_truncation_during_write
Author: Ashish Singh <asingh@cloudera.com>
Reviewers: Jason Gustafson <jason@confluent.io>, Colin P. Mccabe <cmccabe@confluent.io>, Dana Powers <dana.powers@gmail.com>, Gwen Shapira <cshapi@gmail.com>, Grant Henke <granthenke@gmail.com>, Ismael Juma <ismael@juma.me.uk>
Closes#1251 from SinghAsDev/KAFKA-3600
The latter return `Iterable` instead of `Iterator` so that enhanced foreach can be used
in Java.
Author: Ismael Juma <ismael@juma.me.uk>
Reviewers: Jason Gustafson <jason@confluent.io>
Closes#2261 from ijuma/deepEntries-shallowEntries
Tasks that don't have any `StateStore`s wont have a `StandbyTask`, so `createStandbyTask` can return `null`. We need to check for this in `StandbyTaskCreator.createTask(...)`
Also, the checkpointed offsets for `StandbyTask`s are never loaded.
Author: Damian Guy <damian.guy@gmail.com>
Reviewers: Eno Thereska, Matthias J. Sax, Guozhang Wang
Closes#2255 from dguy/kafka-4539
Fix OffsetIndex overflow when replicating a highly compacted topic.
https://issues.apache.org/jira/browse/KAFKA-4451
Author: Michael Schiff <schiff.michael@gmail.com>
Author: Michael Schiff <michael.schiff@tubemogul.com>
Reviewers: Jun Rao <junrao@gmail.com>
Closes#2210 from michaelschiff/bug/4451
Author: Jason Gustafson <jason@confluent.io>
Reviewers: Ismael Juma <ismael@juma.me.uk>, Guozhang Wang <wangguoz@gmail.com>, Jun Rao <junrao@gmail.com>
Closes#2140 from hachikuji/KAFKA4390
Author: Eno Thereska <eno.thereska@gmail.com>
Reviewers: Ismael Juma <ismael@juma.me.uk>, Jason Gustafson <jason@confluent.io>
Closes#2193 from enothereska/KAFKA-4405-prefetch
Improve consumer metric collection by collecting and recording metrics per topic.
Author: Vahid Hashemian <vahidhashemian@us.ibm.com>
Reviewers: Jason Gustafson <jason@confluent.io>
Closes#1684 from vahidhashemian/KAFKA-4000
Collecting socket server metrics during shutdown may throw NullPointerException
Author: Xavier Léauté <xavier@confluent.io>
Reviewers: Ismael Juma <ismael@juma.me.uk>
Closes#2221 from xvrl/fix-metrics-npe-on-shutdown
Also:
* Make all implementations of `Time` thread-safe as they are accessed from multiple threads in some cases.
* Change default implementation of `MockTime` to use two separate variables for `nanoTime` and `currentTimeMillis` as they have different `origins`.
Author: Ismael Juma <ismael@juma.me.uk>
Reviewers: Ewen Cheslack-Postava <ewen@confluent.io>, Shikhar Bhushan <shikhar@confluent.io>, Jason Gustafson <jason@confluent.io>, Eno Thereska <eno.thereska@gmail.com>, Damian Guy <damian.guy@gmail.com>
Closes#2095 from ijuma/kafka-2247-consolidate-time-interfaces
Author: Jason Gustafson <jason@confluent.io>
Reviewers: Ismael Juma <ismael@juma.me.uk>, Guozhang Wang <wangguoz@gmail.com>
Closes#2190 from hachikuji/KAFKA-4469
Author: Antony Stubbs <antony.stubbs@gmail.com>
Reviewers: Eno Thereska <eno.thereska@gmail.com>, Ismael Juma <ismael@juma.me.uk>
Closes#2157 from astubbs/trunk
Process requests received from channels before they were closed. For consumers, wait for coordinator requests to complete before returning from close.
Author: Rajini Sivaram <rajinisivaram@googlemail.com>
Reviewers: Jason Gustafson <jason@confluent.io>, Jun Rao <junrao@gmail.com>, Ismael Juma <ismael@juma.me.uk>
Closes#1836 from rajinisivaram/KAFKA-3703
Author: Mickael Maison <mickael.maison@gmail.com>
Reviewers: Jiangjie Qin <becket.qin@gmail.com>, Jason Gustafson <jason@confluent.io>
Closes#1827 from mimaison/KAFKA-4081
[KAFKA-4284](https://issues.apache.org/jira/browse/KAFKA-4284)
Even though Partitioner has a close method it is not closed when the producer is closed. Serializers, interceptors and metrics are all closed, so partitioners should be closed to.
To be able to use the same mechanism to close the partitioner as the serializers, etc. I had to make the `Partitioner` interface extend `Closeable`. Since this doesn't change the interface that feels ok and should be backwards compatible.
Looking at [KAFKA-2091](https://issues.apache.org/jira/browse/KAFKA-2091) (d6c45c70fb) that introduced the `Partitioner` interface it looks like the intention was that the producer should close the partitioner.
This contribution is my original work and I license the work to the project under the project's open source license.
Author: Theo <theo@iconara.net>
Reviewers: Ewen Cheslack-Postava <ewen@confluent.io>
Closes#2000 from iconara/kafka-4284
I think the Javadoc should describe what happens if wakeup is called and no other thread is currently blocking. This may be important in some cases, e.g. trying to shut down a poll thread, followed by manually committing offsets.
Author: Stig Rohde Døssing <sdo@it-minds.dk>
Reviewers: Jason Gustafson <jason@confluent.io>
Closes#2093 from srdo/minor-expand-wakeup-javadoc
Fixes a bug that inappropriately applies backoff as interval between metadata updates even though the current one is outdated.
Author: Yuto Kawamura <kawamuray.dadada@gmail.com>
Reviewers: Ismael Juma <ismael@juma.me.uk>, Jason Gustafson <jason@confluent.io>
Closes#1707 from kawamuray/KAFKA-4024-metadata-backoff
1. Create a new `ClientMetadata` to collapse `Set<String> consumerMemberIds`, `ClientState<TaskId> state`, and `HostInfo hostInfo`.
2. Stop reusing `stateChangelogTopicToTaskIds` and `internalSourceTopicToTaskIds` to access the (sub-)topology's internal repartition and changelog topics for clarity; also use the source topics num.partitions to set the num.partitions for repartition topics, and clarify to NOT have cycles since otherwise the while loop will fail.
3. `ensure-copartition` at the end to modify the number of partitions for repartition topics if necessary to be equal to other co-partition topics.
4. Refactor `ClientState` as well and update the logic of `TaskAssignor` for clarity as well.
5. Change default `clientId` from `applicationId-suffix` to `applicationId-processId` where `processId` is an UUID to avoid conflicts of clientIds that are from different JVMs, and hence conflicts in metrics.
6. Enforce `assignment` partitions to have the same size, and hence 1-1 mapping to `activeTask` taskIds.
7. Remove the `AssignmentSupplier` class by always construct the `partitionsByHostState` before assigning tasks to consumers within a client.
8. Remove all unnecessary member variables in `StreamPartitionAssignor`.
9. Some other minor fixes on unit tests, e.g. remove `test only` functions with java field reflection.
Author: Guozhang Wang <wangguoz@gmail.com>
Reviewers: Xavier Léauté, Matthias J. Sax, Eno Thereska, Jason Gustafson
Closes#2012 from guozhangwang/K4117-stream-partitionassignro-cleanup
There should be only one cases where these clean-ups have a functional impact: replaced repeated identical logs with a single log for the stale controller epoch case.
The rest should just make the code easier to read and make it a bit less wasteful. I did this exercise because unused variables sometimes mask bugs.
Author: Ismael Juma <ismael@juma.me.uk>
Reviewers: Jason Gustafson <jason@confluent.io>
Closes#1985 from ijuma/remove-unused
Increase timeout in test to avoid transient failures due to long GC or slow machine.
Author: Rajini Sivaram <rajinisivaram@googlemail.com>
Reviewers: Jun Rao <junrao@gmail.com>
Closes#2057 from rajinisivaram/KAFKA-2089
Author: Jason Gustafson <jason@confluent.io>
Reviewers: Guozhang Wang <wangguoz@gmail.com>, Ismael Juma <ismael@juma.me.uk>
Closes#2031 from hachikuji/KAFKA-4303
Author: Konstantine Karantasis <konstantine@confluent.io>
Reviewers: Ismael Juma <ismael@juma.me.uk>, Jason Gustafson <jason@confluent.io>
Closes#1995 from kkonstantine/KAFKA-4254-Update-producers-metadata-before-failing-on-non-existent-partition
And improve readability by adding proper punctuations.
Author: Vahid Hashemian <vahidhashemian@us.ibm.com>
Reviewers: Jason Gustafson <jason@confluent.io>
Closes#2002 from vahidhashemian/doc/fix_typos
Author: Jiangjie Qin <becket.qin@gmail.com>
Reviewers: Ismael Juma <ismael@juma.me.uk>, Jason Gustafson <jason@confluent.io>
Closes#1993 from becketqin/KAFKA-4274
Reopening of https://github.com/apache/kafka/pull/1428
Author: Edoardo Comar <ecomar@uk.ibm.com>
Author: Mickael Maison <mickael.maison@gmail.com>
Reviewers: Grant Henke <granthenke@gmail.com>, Ismael Juma <ismael@juma.me.uk>, Jason Gustafson <jason@confluent.io>
Closes#1908 from edoardocomar/KAFKA-3396
Author: Jason Gustafson <jason@confluent.io>
Reviewers: Vahid Hashemian <vahidhashemian@us.ibm.com>, Jiangjie Qin <becket.qin@gmail.com>, Ismael Juma <ismael@juma.me.uk>
Closes#1936 from hachikuji/KAFKA-3824
Temporarily disable the offset commit (when auto commit is enabled) in the new consumer's `unsubscribe()` method towards a workaround for the issue reported in [KAFKA-3491](https://issues.apache.org/jira/browse/KAFKA-3491).
For now, a call to `unsubscribe()` can be made to reset the offsets in case processing the batch received from the most recent `poll()` is interrupted (due to some exception).
Author: Vahid Hashemian <vahidhashemian@us.ibm.com>
Reviewers: Jason Gustafson <jason@confluent.io>
Closes#1944 from vahidhashemian/KAFKA-4234
Author: Matthias J. Sax <matthias@confluent.io>
Reviewers: Guozhang Wang <wangguoz@gmail.com>, Eno Thereska <eno.thereska@gmail.com>, Ismael Juma <ismael@juma.me.uk>
Closes#1938 from mjsax/kafka-4058-resetIntegartionTest
Author: Jiangjie Qin <becket.qin@gmail.com>
Reviewers: Ismael Juma <ismael@juma.me.uk>, Joel Koshy <jjkoshy.w@gmail.com>, Jason Gustafson <jason@confluent.io>
Closes#1939 from becketqin/KAFKA-4235
Author: Jiangjie Qin <becket.qin@gmail.com>
Reviewers: Ismael Juma <ismael@juma.me.uk>, Jason Gustafson <jason@confluent.io>
Closes#1897 from becketqin/KAFKA-4194
Dong Lin
LoneRifle
MURAKAMI Masahiko
Rajini Sivaram
Jason Gustafson
Ismael Juma
Rekha Joshi
Kamil Szymanski
Ashish Singh
Damian Guy
Michael Schiff
Eno Thereska
Vahid Hashemian
Stig Rohde Døssing
Jeff Widman
Xavier Léauté
Antony Stubbs
Mickael Maison
Theo
Shikhar Bhushan
Yuto Kawamura
Guozhang Wang
radai-rosenblatt
Konstantine Karantasis
Jiangjie Qin
Edoardo Comar
Matthias J. Sax