All the work that we have done to automate and to simplify the coordinator records allows us to simplify the related MessageParsers in DumpLogSegments. They can all share the same based implementation. This is nice because it ensures that we handle all those records similarly.
Reviewers: Chia-Ping Tsai <chia7712@gmail.com>
This patch is a first step towards removing `ReplicaManager#becomeLeaderOrFollower`. It updates the `LocalLeaderEndPointTest` tests.
Reviewers: Christo Lolov <lolovc@amazon.com>, Ismael Juma <ismael@juma.me.uk>
There is a subtle race condition with transactions V2 if a transaction is still completing when checking if we need to add a partition, but it completes when the request reaches the coordinator.
One approach was to remove the verification for TV2 and just check the epoch on write, but a simpler one is to simply return concurrent transactions from the partition leader (before attempting to add the partition). I've done this and added a test for this behavior.
Locally, I reproduced the race but adding a 1 second sleep when handling the WriteTxnMarkersRequest and a 2 second delay before adding the partition to the AddPartitionsToTxnManager. Without this change, the race happened on every second transaction as the first one completed. With this change, the error went away.
As a followup, we may want to clean up some of the code and comments with respect to verification as the code is used by both TV0 + verification and TV2. But that doesn't need to complete for 4.0. This does :)
Reviewers: Jeff Kim <jeff.kim@confluent.io>, Artem Livshits <alivshits@confluent.io>, Calvin Liu <caliu@confluent.io>
While setting Defaults.LogMessageTimestampType to "LogAppendTime", `LogCleanerManagerTest.testLogsUnderCleanupIneligibleForCompaction()` fails with a InvalidTimestampException.
This PR fixes this by regenerating the records instead of previous approach of re-using same records in the test.
Reviewers: Divij Vaidya <diviv@amazon.com>, Kvicii <kvicii.yu@gmail.com>
---------
Co-authored-by: fangxiaobing <fangxiaobing@kuaishou.com>
These methods were previously invoked by ZK components, but we have just removed them.
Reviewers: Ismael Juma <ismael@juma.me.uk>, Chia-Ping Tsai <chia7712@gmail.com>
This patch does a few things:
1) Replace ApiMessageAndVersion by ApiMessage in CoordinatorRecord for the key
2) Leverage the fact that ApiMessage exposes the apiKey. Hence we don't need to specify the key anymore.
Reviewers: Andrew Schofield <aschofield@confluent.io>
This patch removes `ReplicaManager#stopReplicas`. I have ensured that removed unit tests are covered by other existing tests or are updated to use kraft.
Reviewers: Ismael Juma <ismael@juma.me.uk>, Chia-Ping Tsai <chia7712@gmail.com>
Although `MetadataCache`'s `getPartitionReplicaEndpoints` takes a single topic-partition, the `KRaftMetadataCache` implementation iterates over all partitions of the matching topic. This is not necessary and can cause significant performance degradation when the topic has a relatively high number of partitions.
Note that this is not a recent regression - it has been a part of `KRaftMetadataCache` since its creation.
Reviewers: Ismael Juma <ismael@juma.me.uk>, David Jacot <djacot@confluent.io>
This patch updates the transaction coordinator record to use the new coordinator record definition.
Reviewers: Andrew Schofield <aschofield@confluent.io>
Kafka 4.0 will remove support for zk mode and will require conversion to kraft
before upgrading to 4.0. The minimum kraft version is 3.0 (aka 3.0-IV1).
This provides an opportunity to remove exclusively server side protocols versions
that only exist to allow direct upgrades from versions older than 3.0 or that are
used only by zk mode.
Since KRaft became production ready in 3.3, we should consider setting the
baseline to 3.3. But that requires more discussion and it can be done via a
separate change (KAFKA-18601).
Protocol changes:
* Remove RequestHeader v0 (only used by ControlledShutdown v0)
* Remove WriteTxnMarkers v0
* Remove all versions of ControlledShutdown, LeaderAndIsr, StopReplica, UpdateMetadata
In order to remove all versions safely, extend generator to support setting
"versions" to "none". In this case, we no longer generate the `*Data` classes,
but we still reserve the id for the relevant protocol api (so it doesn't get
accidentally used for something else). The protocol documentation is correct
after these changes.
We kept a simplified version of `LeaderAndIsr{Request|Response}` because
it's used by many tests that are still relevant in kraft mode. Once
KAFKA-18486 is done, it may be possible to remove it (I left a comment on
the ticket). Similarly, KAFKA-18487 may make it possible to remove
the introduced `StopReplicaPartitionState` (left a comment on that ticket too).
There are a number of places that were adjusted to include an
`ApiKeys.hasValidVersion` check.
Reviewers: Chia-Ping Tsai <chia7712@gmail.com>
The reason for flakiness is PR #18080 which modifies the linger.ms config from 0 to 5. ClientIdQuotaTest are testing "Low enough quota that a producer sending a small payload in a tight loop should get throttled", thus this config change Influence this test scenario.
This commits uses the older value of 0ms for linger.ms for ClientIdQuotaTest tests.
Reviewers: Ismael Juma <ismael@juma.me.uk>, TaiJuWu <tjwu1217@gmail.com>, Divij Vaidya <diviv@amazon.com>
Since maxOverCleanerThreads does not have a corresponding unit test,
I have added a unit test for it. maintainUncleanablePartitions has been
thoroughly tested in tests, so I simply replaced the old implementation
with the new one.
Reviewers: Ismael Juma <ismael@juma.me.uk>, Chia-Ping Tsai <chia7712@gmail.com>
This was removed during removal of zk code (#18542), but
we should instead convert it to work with kraft.
Reviewers: Ismael Juma <ismael@juma.me.uk>, Chia-Ping Tsai <chia7712@gmail.com>
Currently, each log.append() will add at most 1 index entry, even when the appended data is larger than log.index.interval.bytes. One potential issue is that if a follower restarts after being down for a long time, it may fetch data much bigger than log.index.interval.bytes at a time. This means that fewer index entries are created, which can increase the fetch time from the consumers.
Reviewers: Chia-Ping Tsai <chia7712@gmail.com>, Jun Rao <junrao@gmail.com>
This PR implements the second part of KIP-996 and KAFKA-16164 (tasks KAFKA-16607, KAFKA-17642, KAFKA-17643, KAFKA-17675) which encompass the response handling of PreVotes, addition of new ProspectiveState, update to metrics, and addition of Raft simulation tests.
Voters now transition to ProspectiveState first before CandidateState to prevent unnecessary epoch bumps. Voters in ProspectiveState send PreVotes requests which are Vote requests with PreVote set to true.
Follower grants PreVotes if it has not yet fetched successfully from leader. Leader denies all PreVotes. Unattached, Prospective, Candidate, and Resigned will grant PreVotes if the requesting replica's log is at least as long as theirs. Granted PreVotes are not persisted like standard votes. It is possible for a voter to grant several PreVotes in the same epoch.
The only state which is allowed to transition directly to CandidateState is ProspectiveState. This happens on reception of majority of granted PreVotes or if at least one voter doesn't support PreVote requests.
Prospective will transition to Follower after election loss/timeout if it was already aware of last known leader and the leader's endpoint, or at any point if it discovers the leader.
Prospective will transition to Unattached after election loss/timeout if it does not know the leader endpoints.
After electionTimeout, Resigned now always transitions to Unattached and increases the epoch.
Prospective grants standard votes if it has not already granted a standard vote (no votedKey), has no leaderId, and the recipient's log is current enough
Candidate no longer backs off after election timeout. Candidate still backs off after election loss.
Reviewers: José Armando García Sancio <jsancio@apache.org>
Removed Optional for SharePartitionManager and ClientMetricsManager as zookeeper code is being removed. Also removed asScala and asJava conversion in KafkaApis.handleListClientMetricsResources, moved to java stream.
Reviewers: Mickael Maison <mickael.maison@gmail.com>, Ismael Juma <ismael@juma.me.uk>, Chia-Ping Tsai <chia7712@gmail.com>
The PR implements a way to divide acquired batches into batch size as desirable by client.
The BatchSize is the soft limit and should align the batches in response and cached state in broker at the log batch boundaries.
Reviewers: Andrew Schofield <aschofield@confluent.io>, Abhinav Dixit <adixit@confluent.io>, Jun Rao <junrao@gmail.com>
Fix the issue where producer.commitTransaction under transaction version 2 throws error if no partition or offset is added to transaction. The solution is to avoid sending the endTxnRequest unless producer.send or producer.sendOffsetsToTransaction is triggered.
Reviewers: Justine Olshan <jolshan@confluent.io>
The PR removes dependency of server module on share-coordinator, rather it should be other way. Moved the ShareCoordinatorConfig class from server to share-coordinator.
Reviewers: Ismael Juma <ismael@juma.me.uk>, Chia-Ping Tsai <chia7712@gmail.com>
The REMOTE_LOG_MANAGER_TASK_INTERVAL_MS_PROP in RemoteLogManagerTest is 100 which is too small. If assertions verifyNoMoreInteractions can't run in 100ms, the scheduler will run RLMTask again and the case will fail.
Reviewers: Luke Chen <showuon@gmail.com>
This patch updates the ShareGroupStateMessageParser and OffsetsMessageParser used by the DumpLogSegments command line tool to use the recently introduced json converters for those records. It basically means that new records are automatically supported.
Reviewers: Mickael Maison <mickael.maison@gmail.com>, Chia-Ping Tsai <chia7712@gmail.com>
Apache Kafka 4.0 will only support KRaft and 3.0-IV1 is the minimum version supported by KRaft. So, we can assume that Apache Kafka 4.0 will only communicate with brokers that are 3.0-IV1 or newer.
Note that KRaft was only marked as production-ready in 3.3, so we could go further and set the baseline to 3.3. I think we should have that discussion, but it made sense to start with the non controversial parts.
Reviewers: Jun Rao <junrao@gmail.com>, Chia-Ping Tsai <chia7712@gmail.com>, David Jacot <david.jacot@gmail.com>
To avoid self-fencing in the commit/abort + empty abort scenario, return the concurrent transactions error when we have pending state and do the epoch check second. In this scenario, we will complete the previous transaction before proceeding to the empty abort.
Added a test that failed before the change.
Note -- only the pending state is checked earlier. This is because we don’t return from EndTxn (the first commit) until we already written to the log, transitioned to PrepareX, and have the pending CompleteX state. We don't need to worry about the cases of an EndTxn request coming in with PrepareX without the pending state because that would be an older request and/or retry which are already covered.
Reviewers: Artem Livshits <alivshits@confluent.io>, Jeff Kim <jeff.kim@confluent.io>
The test has become flakier recently and it's easy to reproduce by running the single test (vs
running the the class test suite).
The root cause is that following functions call `RemoteLogMetadataManager#listRemoteLogSegments`.
It returns iterator. If one of function goes through iterator first, another can't get expected result.
I changed `thenReturn` to `thenAnswer` to avoid the issue.
The race is between:
* RLMExpirationTask#cleanupExpiredRemoteLogSegments
* RemoteLogManager#deleteRemoteLogPartition
Reviewers: Ismael Juma <ismael@juma.me.uk>
Signed-off-by: PoAn Yang <payang@apache.org>
Convert v0/v1 record batches to v2 during compaction even if said record batches would be
written with no change otherwise. A few important details:
1. V0 compressed record batch with multiple records is converted into single V2 record batch
2. V0 uncompressed records are converted into single record V2 record batches
3. V0 records are converted to V2 records with timestampType set to `CreateTime` and the
timestamp is `-1`.
4. The `KAFKA-4298` workaround is no longer needed since the conversion to V2 fixes
the issue too.
5. Removed a log warning applicable to consumers older than 0.10.1 - they are no longer
supported.
6. Added back the ability to append records with v0/v1 (for testing only).
7. The creation of the leader epoch cache is no longer optional since the record version
config is effectively always V2.
Add integration tests, these tests existed before #18267 - restored, modified and
extended them.
Reviewers: Jun Rao <jun@confluent.io>
Delete the handlers for LEADER_AND_ISR, STOP_REPLICA, and UPDATE_METADATA. Also, remove the corresponding unit tests in KafkaApisTest.
Reviewers: Chia-Ping Tsai <chia7712@gmail.com>
Remove Apache ZooKeeper from the Apache Kafka build. Also remove commons IO, commons CLI, and netty, which were dependencies we took only because of ZooKeeper.
In order to keep the size of this PR manageable, I did not remove all classes which formerly interfaced with ZK. I just removed the ZK types. Fortunately, Kafka generally wrapped ZK data structures rather than using them directly.
Some classes were pretty entangled with ZK, so it was easier just to stub them out. For ZkNodeChangeNotificationListener.scala, PartitionStateMachine.scala, ReplicaStateMachine.scala, KafkaZkClient.scala, and ZookeeperClient.scala, I replaced all the functions with "throw new UnsupportedOperationException". Since the tests for these classes have been removed, as well as the ZK-based broker code, this should be OK as an incremental step.
Reviewers: Chia-Ping Tsai <chia7712@gmail.com>
Looking at the [history](https://ge.apache.org/scans/tests?search.rootProjectNames=kafka&search.timeZoneId=Europe%2FZurich&tests.container=kafka.api.PlaintextAdminIntegrationTest&tests.test=testConsumerGroups(String%2C%20String)%5B2%5D), I found out that one source of flakiness is due to syncCommit failing with CommitFailedException. We can ignore it and retry on the next iteration.
```
[2025-01-07 10:17:00,783] ERROR [Consumer instanceId=test_instance_id_1, clientId=test_client_id, groupId=test_group_id] OffsetCommit failed for member VfImExrxT3-w_HNJcTkqnw with stale member epoch error. Last epoch sent: 2 (org.apache.kafka.clients.consumer.internals.CommitRequestManager:773)
Exception in thread "Thread-6" org.apache.kafka.clients.consumer.CommitFailedException: OffsetCommit failed with stale member epoch.The member epoch is stale. The member must retry after receiving its updated member epoch via the ConsumerGroupHeartbeat API.
at org.apache.kafka.clients.consumer.internals.CommitRequestManager.commitSyncExceptionForError(CommitRequestManager.java:481)
at org.apache.kafka.clients.consumer.internals.CommitRequestManager.lambda$commitSyncWithRetries$7(CommitRequestManager.java:472)
at java.base/java.util.concurrent.CompletableFuture.uniWhenComplete(CompletableFuture.java:863)
at java.base/java.util.concurrent.CompletableFuture$UniWhenComplete.tryFire(CompletableFuture.java:841)
at java.base/java.util.concurrent.CompletableFuture.postComplete(CompletableFuture.java:510)
at java.base/java.util.concurrent.CompletableFuture.completeExceptionally(CompletableFuture.java:2162)
at org.apache.kafka.clients.consumer.internals.CommitRequestManager$OffsetCommitRequestState.onResponse(CommitRequestManager.java:776)
at org.apache.kafka.clients.consumer.internals.CommitRequestManager$RetriableRequestState.handleClientResponse(CommitRequestManager.java:893)
at org.apache.kafka.clients.consumer.internals.CommitRequestManager$RetriableRequestState.lambda$buildRequestWithResponseHandling$0(CommitRequestManager.java:883)
at java.base/java.util.concurrent.CompletableFuture.uniWhenComplete(CompletableFuture.java:863)
at java.base/java.util.concurrent.CompletableFuture$UniWhenComplete.tryFire(CompletableFuture.java:841)
at java.base/java.util.concurrent.CompletableFuture.postComplete(CompletableFuture.java:510)
at java.base/java.util.concurrent.CompletableFuture.complete(CompletableFuture.java:2147)
at org.apache.kafka.clients.consumer.internals.NetworkClientDelegate$FutureCompletionHandler.onComplete(NetworkClientDelegate.java:433)
at org.apache.kafka.clients.ClientResponse.onComplete(ClientResponse.java:154)
at org.apache.kafka.clients.NetworkClient.completeResponses(NetworkClient.java:669)
at org.apache.kafka.clients.NetworkClient.poll(NetworkClient.java:661)
at org.apache.kafka.clients.consumer.internals.NetworkClientDelegate.poll(NetworkClientDelegate.java:153)
at org.apache.kafka.clients.consumer.internals.ConsumerNetworkThread.runOnce(ConsumerNetworkThread.java:160)
at org.apache.kafka.clients.consumer.internals.ConsumerNetworkThread.run(ConsumerNetworkThread.java:106)
```
Reviewers: Lianet Magrans <lmagrans@confluent.io>
Remove RaftManager.maybeDeleteMetadataLogDir since it was only used during ZK migration, and that code has been removed.
Similarly, remove RaftManagerTest.testKRaftBrokerDoesNotDeleteMetadataLog which tested that function.
Remove AutoTopicCreationManagerTest since it tests the ZK-mode-only AutoTopicReationManager.
Reviewers: Mickael Maison <mickael.maison@gmail.com>, Chia-Ping Tsai <chia7712@gmail.com>
Following https://github.com/apache/kafka/pull/18261, this patch updates the Share Coordinator to use the new record format.
Reviewers: Chia-Ping Tsai <chia7712@gmail.com>, Andrew Schofield <aschofield@confluent.io>
There are times when the controller has a high event processing time, such as during startup, or when creating a topic with many partitions. We can see these processing times in the p99 metric (kafka.controller:type=ControllerEventManager,name=EventQueueProcessingTimeMs), however it's difficult to see exactly which event is causing high processing time.
With DEBUG logs, we see every event along with its processing time. Even with this, it's a bit tedious to find the event with a high processing time.
This PR logs all events which take longer than 2 seconds at ERROR level. This will help identify events that are taking far too long, and which could be disruptive to the operation of the controller. The slow event logging looks like this:
```
[2024-12-20 15:03:39,754] ERROR [QuorumController id=1] Exceptionally slow controller event createTopics took 5240 ms. (org.apache.kafka.controller.EventPerformanceMonitor)
```
Also, every 60 seconds, it logs some event time statistics, including average time, maximum time, and the name of the event which took the longest. This periodic message looks like this:
```
[2024-12-20 15:35:04,798] INFO [QuorumController id=1] In the last 60000 ms period, 333 events were completed, which took an average of 12.34 ms each. The slowest event was handleCommit[baseOffset=0], which took 41.90 ms. (org.apache.kafka.controller.EventPerformanceMonitor)
```
An operator can disable these logs by adding the following to their log4j config:
```
org.apache.kafka.controller.EventPerformanceMonitor=OFF
```
Reviewers: Colin P. McCabe <cmccabe@apache.org>
Remove the flaky annotation from the following tests
* RemoteLogManagerTest#testFetchOffsetByTimestampWithTieredStorageDoesNotFetchIndexWhenExistsLocally
* All the children of BaseConsumerTest#testCoordinatorFailover
* TransactionsTest#testFailureToFenceEpoch
* TransactionsTest#testReadCommittedConsumerShouldNotSeeUndecidedData
* MetricsDuringTopicCreationDeletionTest#testMetricsDuringTopicCreateDelete
* ProduceRequestTest#testProduceWithInvalidTimestamp
Reviewers: Chia-Ping Tsai <chia7712@gmail.com>
A minor refactoring just before merging #18295 introduced a regression and no test failed. Throw the correct exception and add test to verify it. Also refactor the code slightly to make that possible.
Thanks to Chia-Ping for catching the issue.
Reviewers: Chia-Ping Tsai <chia7712@gmail.com>
Clients that support SASL but don't implement KIP-43 (eg Kafka producer/consumer 0.9.0.x) will
fail to connect after this change.
Added unit tests and also manually tested with the console producer 0.9.0.
While testing, I noticed that the logged message when a 0.9.0 Java client is used without sasl is
slightly misleading - fixed that too.
Reviewers: Manikumar Reddy <manikumar.reddy@gmail.com>
This makes it possible to enable request logs for deprecated protocol api versions without enabling it for the rest. Combined with the ability to enable/disable dynamically, it makes it a bit easier to collect the information about deprecated clients that is not available via metrics.
This isn't particularly useful in trunk/4.0 since there are no deprecated api versions in these versions, but it will be useful for older branches. I intend to backport to those branches and add a release note in the backport regarding the change in behavior.
I manually verified that:
1. If the request logger is configured at `INFO` level, only deprecated protocol api versions are logged and they are logged at `INFO` level.
2. If the request logger is configured at `DEBUG` level, all requests are logged but the log level is `INFO` for deprecated protocol api versions and `DEBUG` for the rest.
3. If the request logger is configured at `WARN` level (the default), no requests are logged.
Reviewers: Chia-Ping Tsai <chia7712@gmail.com>
Added transaction version 2 to some of the system tests. Also marking TV2 as production ready.
Also fixes the defaultVersion test.
Reviewers: Jun Rao <jun@confluent.io>
Included in this change:
1. Remove deprecated protocol api versions from json files.
3. Remove fields that are no longer used from json files (affects ListOffsets, OffsetCommit, DescribeConfigs).
4. Remove record down-conversion support from KafkaApis.
5. No longer return `Errors.UNSUPPORTED_COMPRESSION_TYPE` on the fetch path[1].
6. Deprecate `TopicConfig. MESSAGE_DOWNCONVERSION_ENABLE_CONFIG` and made the relevant
configs (`message.downconversion.enable` and `log.message.downcoversion.enable`) no-ops since
down-conversion is no longer supported. It was an oversight not to deprecate this via KIP-724.
7. Fix `shouldRetainsBufferReference` to handle null request schemas for a given version.
8. Simplify producer logic since it only supports the v2 record format now.
9. Fix tests so they don't exercise protocol api versions that have been removed.
10. Add upgrade note.
Testing:
1. System tests have a lot of failures, but those tests fail for trunk too and I didn't see any issues specific to this change - it's hard to be sure given the number of failing tests, but let's not block on that given the other testing that has been done (see below).
3. Java producers and consumers with version 0.9-0.10.1 don't have api versions support and hence they fail in an ungraceful manner: the broker disconnects and the clients reconnect until the relevant timeout is triggered.
4. Same thing seems to happen for the console producer 0.10.2 although it's unclear why since api versions should be supported. I will look into this separately, it's unlikely to be related to this PR.
5. Console consumer 0.10.2 fails with the expected error and a reasonable message[2].
6. Console producer and consumer 0.11.0 works fine, newer versions should naturally also work fine.
7. kcat 1.5.0 (based on librdkafka 1.1.0) produce and consume fail with a reasonable message[3][4].
8. kcat 1.6.0-1.7.0 (based on librdkafka 1.5.0 and 1.7.0 respectively) consume fails with a reasonable message[5].
9. kcat 1.6.0-1.7.0 produce works fine.
10. kcat 1.7.1 (based on librdkafka 1.8.2) works fine for consumer and produce.
11. confluent-go-client (librdkafka based) 1.8.2 works fine for consumer and produce.
12. I will test more clients, but I don't think we need to block the PR on that.
Note that this also completes part of KIP-724: produce v2 and lower as well as fetch v3 and lower are no longer supported.
Future PRs will remove conditional code that is no longer needed (some of that has been done in KafkaApis,
but only what was required due to the schema changes). We can probably do that in master only as it does
not change behavior.
Note that I did not touch `ignorable` fields even though some of them could have been
changed. The reasoning is that this could result in incompatible changes for clients
that use new protocol versions without setting such fields _if_ we don't manually
validate their presence. I will file a JIRA ticket to look into this carefully for each
case (i.e. if we do validate their presence for the appropriate versions, we can
set them to ignorable=false in the json file).
[1] We would return this error if a fetch < v10 was used and the compression topic config was set
to zstd, but we would not do the same for the case where zstd was compressed at the producer
level (the most common case). Since there is no efficient way to do the check for the common
case, I made it consistent for both by having no checks.
[2] ```org.apache.kafka.common.errors.UnsupportedVersionException: The broker is too new to support JOIN_GROUP version 1```
[3]```METADATA|rdkafka#producer-1| [thrd:main]: localhost:9092/bootstrap: Metadata request failed: connected: Local: Required feature not supported by broker (0ms): Permanent```
[4]```METADATA|rdkafka#consumer-1| [thrd:main]: localhost:9092/bootstrap: Metadata request failed: connected: Local: Required feature not supported by broker (0ms): Permanent```
[5] `ERROR: Topic test-topic [0] error: Failed to query logical offset END: Local: Required feature not supported by broker`
Reviewers: David Arthur <mumrah@gmail.com>
Librdkafka totally breaks if produce v3 is removed - it starts sending records with record format v0.
These api versions have to be undeprecated - KIP-896 has been updated.
Reviewers: David Arthur <mumrah@gmail.com>
This is just a mechanical change to make prepareTransitionTo method use named parameters instead of positional parameters.
Reviewers: Justine Olshan <jolshan@confluent.io>, Ritika Reddy <rreddy@confluent.io>
This patch is the first one in a series to improve how coordinator records are managed. It focuses on making coordinator records first class citizen in the generator.
* Introduce `coordinator-key` and `coordinator-value` in the schema;
* Introduce `apiKey` for those. This is done to avoid relying on the version to determine the type.
* It also allows the generator to enforce some rules: the key cannot use flexible versions, the key must have a single version `0`, there must be a key and a value for a given api key, etc.
* It generates an enum with all the coordinator record types. This is pretty handy in the code.
The patch also updates the group coordinators to use those.
Reviewers: Jeff Kim <jeff.kim@confluent.io>, Andrew Schofield <aschofield@confluent.io>
A lot of these tests assumed that the commit/abort happened immediately. Spoiler alert -- it does not.
For some I ensure that the first send of the next transaction is successful before grabbing the epoch. I also loosened some checks since we don't need to guarantee the exact epoch.
I went back and forth with completely deleting testBumpTransactionalEpochWithTV2Enabled since we don't have client side epoch bumps with V2 (which is what the test was originally testing), but I opted to keep it to just confirm the epoch on each transaction -- even in the timeout scenario.
Reviewers: Calvin Liu <caliu@confluent.io>, Artem Livshits <alivshits@confluent.io>, Jeff Kim <jeff.kim@confluent.io>, David Mao <dmao@confluent.io>
When inter.broker.listener is explicitly set, validate that it is not in the set of controller.listener.names.
Reviewers: Colin P. McCabe <cmccabe@apache.org>, David Arthur <mumrah@gmail.com>
Adds a new RPC StreamsGroupDescribe that returns, given the group ID, all metadata related to the streams group, such as
- The topology metadata of the group.
- The topology epoch of the group.
- The latest member metadata that each member provided through the StreamsGroupHeartbeat API.
- The current target assignment generated by the assignor.
- This just adds the JSON as defined in KIP-1071, together with some plumbing.
Reviewers: Bill Bejeck <bbejeck@gmail.com>
The StreamsGroupHeartbeat API is the new core API used by streams application to form a group. The API allows members to initialize a topology, advertise their state, and their owned tasks. The group coordinator uses it to assign/revoke tasks to/from members. This API is also used as a liveness check.
This change adds the JSON definition of the RPC, as defined in KIP-1071.
Reviewers: Bruno Cadonna <cadonna@apache.org>
The issue has been fixed by https://issues.apache.org/jira/browse/KAFKA-18188. We can re-enable the test with the CONSUMER protocol.
Reviewers: Lianet Magrans <lmagrans@confluent.io>, Andrew Schofield <aschofield@confluent.io>
When a static member rejoins the group, the group state is rewritten to the partition in order to persist the change. If the write fails, the change is reverted. However, this is done without acquiring the group lock.
This is only try in the old group coordinator. The new one does not suffer from this issue.
Reviewers: Jeff Kim <jeff.kim@confluent.io>
Kafka consumer supports auto.offset.reset config option, which is used when there is no initial offset in Kafka (or) if the current offset does not exist any more on the server. This config currently supports earliest/latest/none options. Currently consumer resets might force applications to reprocess large amounts of data from earlier offsets. With infinite storage, its beneficial to have a duration based offset reset strategy. This will allow applications to consume/initialise from a fixed duration when there is no initial offset in Kafka.
As part of KIP-932, we are adding support for share consumer groups. Share consumer groups supports dynamic group configuration property share.auto.offset.reset. This is used to set the initial Share-Partition Start Offset (SPSO) based on the share.auto.offset.reset configuration. Currently share.auto.offset.reset supports earliest and latest options to automatically reset the offset
Similar to the Kafka Consumer, we will add support for by_duration: config value for share.auto.offset.reset.
Reviewers: Andrew Schofield <aschofield@confluent.io>, Apoorv Mittal <apoorvmittal10@gmail.com>
- Disallow configuring -1 for copier and expiration thread pools dynamically
Co-authored-by: Peter Lee <peterxcli@gmail.com>
Reviewers: Peter Lee <peterxcli@gmail.com>, Satish Duggana <satishd@apache.org>
We have observed the below errors in some cluster:
Uncaught exception in scheduled task 'handleTxnCompletion-902667' exception.message:Trying to complete a transactional offset commit for producerId *** and groupId *** even though the offset commit record itself hasn't been appended to the log.
When a transaction is completed, the transaction coordinator sends a WriteTxnMarkers request to all the partitions involved in the transaction to write the markers to them. When the broker receives it, it writes the markers and if markers are written to the __consumer_offsets partitions, it informs the group coordinator that it can materialize the pending transactional offsets in its main cache. The group coordinator does this asynchronously since Apache Kafka 2.0, see this patch.
The above error appends when the asynchronous operation is executed by the scheduler and the operation finds that there are pending transactional offsets that were not written yet. How come?
There is actually an issue is the steps described above. The group coordinator does not wait until the asynchronous operation completes to return to the api layer. Hence the WriteTxnMarkers response may be send back to the transaction coordinator before the async operation is actually completed. Hence it is possible that the next transactional produce to be started also before the operation is completed too. This could explain why the group coordinator has pending transactional offsets that are not written yet.
There is a similar issue when the transaction is aborted. However on this path, we don't have any checks to verify whether all the pending transactional offsets have been written or not so we don't see any errors in our logs. Due to the same race condition, it is possible to actually remove the wrong pending transactional offsets.
PS: The new group coordinator is not impacted by this bug.
Reviewers: Justine Olshan <jolshan@confluent.io>
This pull request replaces Log4j with Log4j2 across the entire project, including dependencies, configurations, and code. The notable changes are listed below:
1. Introduce Log4j2 Instead of Log4j
2. Change Configuration File Format from Properties to YAML
3. Adds warnings to notify users if they are still using Log4j properties, encouraging them to transition to Log4j2 configurations
Co-authored-by: Lee Dongjin <dongjin@apache.org>
Reviewers: Luke Chen <showuon@gmail.com>, Mickael Maison <mickael.maison@gmail.com>, Chia-Ping Tsai <chia7712@gmail.com>
mplementation of KIP-1073: Return fenced brokers in DescribeCluster response.
Add new unit and integration tests for describeCluster.
Reviewers: Luke Chen <showuon@gmail.com>
In this PR, we've added a class ShareCoordinatorOffsetsManager, which tracks the last redundant offset for each share group state topic partition. We have also added a periodic timer job in ShareCoordinatorService which queries for the redundant offset at regular intervals and if a valid value is found, issues the deleteRecords call to the ReplicaManager via the PartitionWriter. In this way the size of the partitions is kept manageable.
Reviewers: Jun Rao <junrao@gmail.com>, David Jacot <djacot@confluent.io>, Andrew Schofield <aschofield@confluent.io>
The LeaveGroup API is used by the admin client to remove static members or remove all members from the group. The latter does not work because the API does not allow removing a member using the CONSUMER protocol by member id. Moreover, the response should only include the member id if the member id was included in the request. This patch fixes both issues.
Reviewers: Dongnuo Lyu <dlyu@confluent.io>, Christo Lolov <lolovc@amazon.com>, Jeff Kim <jeff.kim@confluent.io>
We can better keep track of which transactions use TV_2 by storing this information in the clientTransactionVersion field and persisting it across state transitions. Also updated some logging and equality code to include this information.
Added a test to ensure version persists. There aren't many TV2 transitions that don't specify TV, but I did test the InitProducerId + epoch overflow case.
Reviewers: Artem Livshits <alivshits@confluent.io>, Jeff Kim <jeff.kim@confluent.io>
* Add fields `groupEpoch` and `targetAssignmentEpoch` to `ConsumerGroupDescription.java`.
* Add fields `memberEpoch` and `upgraded` to `MemberDescription.java`.
* Add assertion to `PlaintextAdminIntegrationTest#testDescribeClassicGroups` to make sure member in classic group returns `upgraded` as `Optional.empty`.
* Add new case `testConsumerGroupWithMemberMigration` to `PlaintextAdminIntegrationTest` to make sure migration member has correct `upgraded` value. Add assertion for `groupEpoch`, `targetAssignmentEpoch`, `memberEpoch` as well.
Reviewers: David Jacot <djacot@confluent.io>
Signed-off-by: PoAn Yang <payang@apache.org>
Test lacked the proper configuration for the offset topic replication. As a result, when the follower was shut down, the coordinator did not failover properly.
Reviewers: TaiJuWu <tjwu1217@gmail.com>, David Jacot <djacot@confluent.io>
When KafkaRaftClient receives a request to create a snapshot with end offset that is not aligned to a batch boundary, do not create a misaligned snapshot and instead log at info level and throw an IllegalArgumentException.
Checking that the end offset is at a batch boundary is performed by reading the log at snapshotId.offset() and checking whether the offset in question is the base offset of the returned batch.
Reviewers: José Armando García Sancio <jsancio@apache.org>
Disallow upgrades from classic groups to consumer groups when any member's assignment has non-empty userData.
Reviewers: David Jacot <djacot@confluent.io>
Removes the client side AddPartitionsToTxn/AddOffsetsToTxn calls so that the partition is implicitly added as part of KIP-890 part 2.
This change also requires updating the valid state transitions. The client side can not know for certain if a partition has been added server side when the request times out (partial completion). Thus for TV2, the transition to PrepareAbort is now valid for Empty, CompleteCommit, and CompleteAbort.
For readability, the V1 and V2 endTransaction methods have been separated.
Reviewers: Artem Livshits <alivshits@confluent.io>, Justine Olshan <jolshan@confluent.io>, Ritika Reddy <rreddy@confluent.io>
Ken Huang
David Jacot
TengYao Chi
Logan Zhu
TaiJuWu
Sushant Mahajan
Justine Olshan
Lianet Magrans
Xiaobing Fang
Chung, Ming-Yen
Dimitar Dimitrov
Ismael Juma
Sanskar Jhajharia
PoAn Yang
Alyssa Huang
Kirk True
Mickael Maison
Jason Taylor
Dmitry Werner
Apoorv Mittal
Kuan-Po Tseng
Ted Yan
mingdaoy
Peter Lee
Abhinav Dixit
Xuan-Zhang Gong
Colin Patrick McCabe
Andrew Schofield
Jhen-Yung Hsu
NICOLAS GUYOMAR
David Arthur
Zhihong Yu
xijiu
Artem Livshits
kevin-wu24
Lucas Brutschy
Nick Guo
David Jacot
ShivsundarR
Kamal Chandraprakash
Gantigmaa Selenge
Oleksandr K.
Sean Quah
Calvin Liu