When debugging KafkaConsumer production issues, it's pretty
useful to have log entries related to seeking and committed
offset retrieval enabled by default. These are currently present,
but only when debug logging is enabled. Change them to `info`.
Also included a minor code simplication and a slight improvement
to an exception message.
Reviewers: Jason Gustafson <jason@confluent.io>
It used to preallocate an array of responses and then complete each response from the original collection sequentially. The problem was that the original collection could have been modified (another thread completing the response) while this was hapenning
Avoid unnecessary lock acquire when KafkaConsumer commits offsets.
Reviewers: Viktor Somogyi-Vass <viktorsomogyi@gmail.com>, Guozhang Wang <wangguoz@gmail.com>
As part of KIP-320, we should add the expected leader epoch to Fetch and ListOffsets requests. This will allow us ultimately to detect log truncation.
Reviewers: Jason Gustafson <jason@confluent.io>
- Include more detail in the client log message if the disconnection happens
during authentication.
- Include exception message in the Selector info entry when authentication
fails and unwrap `DelayedResponseAuthenticationException`.
- Remove duplicate debug log on authentication failure in the Selector.
Empty username or password would result in the "expected 3 tokens"
error instead of "username not specified" or "password not specified".
Reviewers: Rajini Sivaram <rajinisivaram@googlemail.com>
JUnit 4.13 fixes the issue where `Category` and `Parameterized` annotations
could not be used together. It also deprecates `ExpectedException` and
`assertThat`. Given this, we:
- Replace `ExpectedException` with the newly introduced `assertThrows`.
- Replace `Assert.assertThat` with `MatcherAssert.assertThat`.
- Annotate `AbstractLogCleanerIntegrationTest` with `IntegrationTest` category.
Reviewers: Ewen Cheslack-Postava <ewen@confluent.io>, David Arthur <mumrah@gmail.com>
Don't return error messages from `SaslException` to clients. Error messages to be returned to clients to aid debugging must be thrown as AuthenticationExceptions. This is a fix for a regression from KAFKA-7352.
Reviewers: Ron Dagostino <rndgstn@gmail.com>, Ismael Juma <ismael@juma.me.uk
This helps narrow down the specific broker they came from when debugging
ACL propagation issues.
Reviewers: Vahid Hashemian <vahid.hashemian@gmail.com>, Ismael Juma <ismael@juma.me.uk>
When an older message format is in use, we should disable the leader epoch cache so that we resort to truncation by high watermark. Previously we updated the cache for all versions when a broker became leader for a partition. This can cause large and unnecessary truncations after leader changes because we relied on the presence of _any_ cached epoch in order to tell whether to use the improved truncation logic possible with the OffsetsForLeaderEpoch API.
Reviewers: Stanislav Kozlovski <stanislav_kozlovski@outlook.com>, Viktor Somogyi-Vass <viktorsomogyi@gmail.com>, Jun Rao <junrao@gmail.com>
Replaces original loginContext if login fails in the refresh thread to ensure that the refresh thread is left in a clean state when there are exceptions while connecting to an OAuth server. Also makes client callback handler more robust by using the token with the longest remaining time for expiry instead of throwing an exception if multiple tokens are found.
Reviewers: Rajini Sivaram <rajinisivaram@googlemail.com>
This patch introduces a new config - "group.max.size", which caps the maximum size any group can reach. It has a default value of Int.MAX_VALUE. Once a group is of the maximum size, subsequent JoinGroup requests receive a MAX_SIZE_REACHED error.
In the case where the config is changed and a Coordinator broker with the new config loads an old group that is over the threshold, members are kicked out of the group and a rebalance is forced.
Reviewers: Vahid Hashemian <vahid.hashemian@gmail.com>, Boyang Chen <bchen11@outlook.com>, Gwen Shapira <cshapi@gmail.com>, Jason Gustafson <jason@confluent.io>
See also KIP-183.
This implements the following algorithm:
AdminClient sends ElectPreferredLeadersRequest.
KafakApis receives ElectPreferredLeadersRequest and delegates to
ReplicaManager.electPreferredLeaders()
ReplicaManager delegates to KafkaController.electPreferredLeaders()
KafkaController adds a PreferredReplicaLeaderElection to the EventManager,
ReplicaManager.electPreferredLeaders()'s callback uses the
delayedElectPreferredReplicasPurgatory to wait for the results of the
election to appear in the metadata cache. If there are no results
because of errors, or because the preferred leaders are already leading
the partitions then a response is returned immediately.
In the EventManager work thread the preferred leader is elected as follows:
The EventManager runs PreferredReplicaLeaderElection.process()
process() calls KafkaController.onPreferredReplicaElectionWithResults()
KafkaController.onPreferredReplicaElectionWithResults()
calls the PartitionStateMachine.handleStateChangesWithResults() to
perform the election (asynchronously the PSM will send LeaderAndIsrRequest
to the new and old leaders and UpdateMetadataRequest to all brokers)
then invokes the callback.
Reviewers: Colin P. McCabe <cmccabe@apache.org>, Jun Rao <junrao@gmail.com>
The problem is that the sequence number is an Int and should wrap around when it reaches the Int.MaxValue. The bug here is it doesn't wrap around and become negative and raises an error.
Reviewers: Jason Gustafson <jason@confluent.io>
This patch fixes a few overflow issues with wrapping sequence numbers in the broker's producer state tracking.
Reviewers: Jason Gustafson <jason@confluent.io>
The presence of the buildSrc subproject is causing problems when we try
to run installAll, jarAll, and the other "all" targets. It's easier
just to make the generator code a regular subproject and use the
JavaExec gradle task to run the code. This also makes it more
straightforward to run the generator unit tests.
Reviewers: David Arthur <mumrah@gmail.com>, Ismael Juma <ismael@juma.me.uk>
Co-authored-by: Colin P. Mccabe <cmccabe@confluent.io>
Co-authored-by: Stanislav Kozlovski <stanislav_kozlovski@outlook.com>
* Join ssl principal mapping rules correctly before evaluating.
Java properties splits the configuration array on commas, and that leads to rules containing commas being split before being evaluated. This commit adds a code change to re-join those strings into full rules before evaluating them. The function assumes every rule is either DEFAULT or begins with the prefix RULE:
The Javadoc is using Properties.put which should never be used because it allows putting non-strings into a Properties object which is designed to only handle strings.
Two other minor fixes so the examples actually work
Track the last seen partition epoch in the Metadata class. When handling metadata updates, check that the partition info being received is for the last seen epoch or a newer one. This prevents stale metadata from being loaded into the client.
Reviewers: Jason Gustafson <jason@confluent.io>
This patch adds a framework to automatically generate the request/response classes for Kafka's protocol. The code will be updated to use the generated classes in follow-up patches. Below is a brief summary of the included components:
**buildSrc/src**
The message generator code is here. This code is automatically re-run by gradle when one of the schema files changes. The entire directory is processed at once to minimize the number of times we have to start a new JVM. We use Jackson to translate the JSON files into Java objects.
**clients/src/main/java/org/apache/kafka/common/protocol/Message.java**
This is the interface implemented by all automatically generated messages.
**clients/src/main/java/org/apache/kafka/common/protocol/MessageUtil.java**
Some utility functions used by the generated message code.
**clients/src/main/java/org/apache/kafka/common/protocol/Readable.java, Writable.java, ByteBufferAccessor.java**
The generated message code uses these classes for writing to a buffer.
**clients/src/main/message/README.md**
This README file explains how the JSON schemas work.
**clients/src/main/message/\*.json**
The JSON files in this directory implement every supported version of every Kafka API. The unit tests automatically validate that the generated schemas match the hand-written schemas in our code. Additionally, there are some things like request and response headers that have schemas here.
**clients/src/main/java/org/apache/kafka/common/utils/ImplicitLinkedHashSet.java**
I added an optimization here for empty sets. This is useful here because I want all messages to start with empty sets by default prior to being loaded with data. This is similar to the "empty list" optimizations in the `java.util.ArrayList` class.
Reviewers: Stanislav Kozlovski <stanislav_kozlovski@outlook.com>, Ismael Juma <ismael@juma.me.uk>, Bob Barrett <bob.barrett@outlook.com>, Jason Gustafson <jason@confluent.io>
* Update KafkaAdminClient#describeTopics to throw UnknownTopicOrPartitionException.
* Remove unused method: WorkerUtils#getMatchingTopicPartitions.
* Add some JavaDoc.
Reviewed-by: Colin P. McCabe <cmccabe@apache.org>, Ryanne Dolan <ryannedolan@gmail.com>
There is a race condition in ReplicaFetcherThread, where we can update PartitionFetchState with the new leader epoch (same leader) before handling the OffsetsForLeaderEpoch response with FENCED_LEADER_EPOCH error which causes removing partition from partitionStates, which in turn causes no fetching until the next LeaderAndIsr.
This patch adds logic to ensure that the leader epoch doesn't change while an OffsetsForLeaderEpoch request is in flight (which could happen with back-to-back leader elections). If it has changed, we ignore the response.
Also added toString() implementation to PartitionData, because some log messages did not show useful info which I found while investigating the above system test failure.
Reviewers: Ismael Juma <ismael@juma.me.uk>, Jason Gustafson <jason@confluent.io>
Use `info` for failed authentications and `debug` for successful ones.
Reviewers: Rajini Sivaram <rajinisivaram@googlemail.com>, Ismael Juma <ismael@juma.me.uk>
This patch changes the behavior of KafkaProducer.waitOnMetadata to wait up to max.block.ms when the partition specified in the produce request is out of the range of partitions present in the metadata. This improves the user experience in the case when partitions are added to a topic and a client attempts to produce to one of the new partitions before the metadata has propagated to the brokers. Tested with unit tests.
Reviewers: Arjun Satish <arjun@confluent.io>, Jason Gustafson <jason@confluent.io>
Lookup client host name after every full iteration through the addresses returned.
Reviewers: Loïc Monney <loicmonney@github.com>, Edoardo Comar <ecomar@uk.ibm.com>, Rajini Sivaram <rajinisivaram@googlemail.com>
1. The retry loop of the InternalTopicManager would just be: a) describe topics, and exclude those which already exist with the right num.partitions, b) for the remaining topics, try to create them. Remove any inner loops.
2. In CreateTopicResponse and MetadataResponse (for describe topic), handle the special error code of TopicExist and UnknownTopicOrPartition in order to retry in the next loop.
3. Do not handle TimeoutException since it should already been handled inside AdminClient.
Add corresponding unit tests for a) topic marked for deletion but not complete yet, in which case metadata response would not contain this topic, but create topic would return error TopicExists; b) request keep getting timed out.
Reviewers: Matthias J. Sax <matthias@confluent.io>
This pull request replaces HashMap with LinkedHashMap to guarantee ordering of metrics tags.
Reviewers: Ismael Juma <ismael@juma.me.uk>, Guozhang Wang <guozhang@confluent.io>, John Roesler <vvcephei@users.noreply.github.com>
Check `running` in `Sender.maybeWaitForProducerId` to ensure that the producer can be closed while awaiting initialization of the producerId.
Reviewers: Jason Gustafson <jason@confluent.io>
Replace `channel` by `fileRecords` in potentially thrown KafkaException
descriptions when loading/writing `FileChannelRecordBatch`. This makes exception
messages more readable (channel only shows an object hashcode, fileRecords shows
the path of the file being read and start/end positions in the file).
Reviewers: Jason Gustafson <jason@confluent.io>
After a recent leader election, the leaders high-water mark might lag behind the offset at the beginning of the new epoch (as well as the previous leader's HW). This can lead to offsets going backwards from a client perspective, which is confusing and leads to strange behavior in some clients.
This change causes Partition#fetchOffsetForTimestamp to throw an exception to indicate the offsets are not yet available from the leader. For new clients, a new OFFSET_NOT_AVAILABLE error is added. For existing clients, a LEADER_NOT_AVAILABLE is thrown.
This is an implementation of [KIP-207](https://cwiki.apache.org/confluence/display/KAFKA/KIP-207%3A+Offsets+returned+by+ListOffsetsResponse+should+be+monotonically+increasing+even+during+a+partition+leader+change).
Reviewers: Colin Patrick McCabe <colin@cmccabe.xyz>, Dhruvil Shah <dhruvil@confluent.io>, Jason Gustafson <jason@confluent.io>
Ensure that channel and selection keys are removed from `Selector` collections before propagating connect exceptions. They are currently cleared on the next `poll()`, but we can't ensure that callers (NetworkClient for example) wont try to connect again before the next `poll` and hence we should clear the collections before re-throwing exceptions from `connect()`.
Reviewers: Jason Gustafson <jason@confluent.io>
The example in the producer's javadoc contained an inconsistent value for `delivery.timeout.ms`. This patch removes the inconsistent config and several unnecessary overrides in order to simplify the example.
Reviewers: huxi <huxi_2b@hotmail.com>, Ismael Juma <ismael@juma.me.uk>, Jason Gustafson <jason@confluent.io>
Ismael Juma
Stanislav Kozlovski
lambdaliu
David Arthur
Mickael Maison
Rajini Sivaram
Radai Rosenblatt
Jason Gustafson
Ron Dagostino
Mark Cho
Colin Patrick McCabe
Boyang Chen
Gardner Vickers
Tom Bentley
mingaliu
Lee Dongjin
ryannatesmith
Lars Francke
Dong Lin
Chia-Ping Tsai
Anna Povzner
Bob Barrett
hackerwin7
Guozhang Wang
lambdaliu
layfe
Flavien Raynaud
Viktor Somogyi
Matthias Wessendorf
Satish Duggana