https://github.com/apache/kafka/pull/4356 added `batch.size` config property to `FileStreamSourceConnector` but the property was added as required without a default in config definition (`ConfigDef`). This results in validation error during connector startup.
Unit tests were added for both `FileStreamSourceConnector` and `FileStreamSinkConnector` to avoid such issues in the future.
Reviewers: Randall Hauch <rhauch@gmail.com>, Jason Gustafson <jason@confluent.io>
The log cleaner should not naively remove the partition from in progress map without checking the partition state. This may cause the other thread calling `LogCleanerManager.abortAndPauseCleaning()` to hang indefinitely.
Enable deep-iteration option when print-data-log is enabled in DumpLogSegments. Otherwise data is not printed.
Reviewers: Jason Gustafson <jason@confluent.io>, Ismael Juma <ismael@juma.me.uk>
Avoid loosing log/stdout/stderr files on restart
Reenables tests
Author: Matthias J. Sax <matthias@confluent.io>
Reviewers: Guozhang Wang <guozhang@confluent.io>, Bill Bejeck <bill@confluent.io>
Allows input data to be read from a file and removes .toLowerCase in word count stream
Author: Filipe Agapito <filipe.agapito@gmail.com>
Reviewers: Ted Yu <yuzhihong@gmail.com>, Matthias J. Sax <matthias@confluent.io>
Make it clear in the docs that the rebalance listener is only invoked during an active call to `poll()`. Plus a few additional doc cleanups.
Reviewers: Ismael Juma <ismael@juma.me.uk>, Jason Gustafson <jason@confluent.io>
Currently we hold onto all Records references in a multi-partition fetch response until the full response has completed. This can be a problem when the records have been down-converted since they will be occupying a (potentially large) chunk of memory. This patch changes the behavior in MultiSend so that once a Send is completed, we no longer keep a reference to it, which will allow the Records objects to be freed sooner.
I have added a simple unit test to verify that sends are removed as the MultiSend progresses.
Reviewers: Ismael Juma <ismael@juma.me.uk>
`Node` is immutable so this is safe.
With 100 brokers, 150 topics and 350 partitions, `HashSet.contains` in `RecordAccumulator.ready` took about 40% of the application time. It
is caused by re-calculating a hash code of a leader (Node instance) for
every batch entry. Caching the hashCode reduced the time of
`HashSet.contains` in `RecordAccumulator.ready` to ~2%. The
measurements were taken with Flight Recorder.
Reviewers: Rajini Sivaram <rajinisivaram@googlemail.com>, Ted Yu <yuzhihong@gmail.com>, Ismael Juma <ismael@juma.me.uk>
ZooKeeperClient acquires initializationLock#writeLock to establish a new connection while processing session expiry WatchEvent. ZooKeeperClient#handleRequests acquires initializationLock#readLock, allowing multiple batches of requests to be processed concurrently, but preventing reconnections while processing requests. At the moment, handleRequests holds onto the readLock throughout the method, even while waiting for responses and inflight requests to complete. But responses cannot be delivered if event thread is blocked on the writeLock to process session expiry event. This results in a deadlock. During broker shutdown, the shutdown thread is also blocked since it needs the readLock to perform ZooKeeperClient#unregisterStateChangeHandler, which cannot be acquired if a session expiry had occurred earlier since this thread gets queued behind the event handler thread waiting for writeLock.
This commit reduces locking in ZooKeeperClient#handleRequests to just the non-blocking send, so that session expiry handling doesn't get blocked when a send is blocked waiting for responses. Also moves session expiry handling to a separate thread so that Kafka controller doesn't block the event handler thread when processing session expiry.
Avoids double initialization of resuming tasks
Removes race condition in StreamThreadTest plus code cleanup
Author: Matthias J. Sax <matthias@confluent.io>
Reviewers: Bill Bejeck <bill@confluent.io>, Guozhang Wang <guozhang@confluent.io>
ProducerBatch retains references to MemoryRecordsBuilder and cannot be freed until acks are received. Removing references to buffers used for compression after records are built will enable these to be garbage collected sooner, reducing the risk of OOM.
Reviewers: Ismael Juma <ismael@juma.me.uk>, Jason Gustafson <jason@confluent.io>, Lothsahn <Lothsahn@gmail.com>
…ateMetadataRequestForBrokers
*More detailed description of your change,
if necessary. The PR title and PR message become
the squashed commit message, so use a separate
comment to ping reviewers.*
*Summary of testing strategy (including rationale)
for the feature or bug fix. Unit and/or integration
tests are expected for any behaviour change and
system tests should be considered for larger changes.*
Author: Lucas Wang <luwang@linkedin.com>
Reviewers: Jun Rao <junrao@gmail.com>
Closes#4472 from gitlw/improving_addUpdateMetadataRequestForBrokers
This fixes two alerts flagged on lgtm.com for Apache Kafka.
This dead code alert where InvalidTypeIdException indirectly extends JsonMappingException. The flagged condition with the type test appears after the type test for the latter and thus makes its body dead. I opted to change the order of the tests. Please let me know if this is the intended behavior.
The second commit addresses this out-of-bounds alert.
More alerts can be found here. Note that my colleague Aditya Sharad addressed some of those in the now outdated #2939.
Reviewers: Matthias J. Sax <matthias@confluent.io>, Rajini Sivaram <rajinisivaram@googlemail.com>
This is a small change to parallelize plugin scanning. This may help in some environments where otherwise plugin scanning is slow.
Author: Robert Yokota <rayokota@gmail.com>
Reviewers: Konstantine Karantasis <konstantine@confluent.io>, Randall Hauch <rhauch@gmail.com>, Ewen Cheslack-Postava <ewen@confluent.io>
Closes#4561 from rayokota/K6503-improve-plugin-scanning
Added a second check for race condition where store changelog topic updated during restore, but not if a KTable changelog topic. This will be tricky to test, but I wanted to push the PR to get feedback on the approach.
Reviewers: Guozhang Wang <wangguoz@gmail.com>, Matthias J. Sax <matthias@confluent.io>
Prior to this patch, the consumer always blocks in poll() if there are any partitions which are awaiting their initial positions. This behavior was inconsistent with normal fetch behavior since we allow fetching on available partitions even if one or more of the assigned partitions becomes unavailable _after_ initial offset lookup. With this patch, the consumer will do offset resets asynchronously, which allows other partitions to make progress even if the initial positions for some partitions cannot be found.
I have added several new unit tests in `FetcherTest` and `KafkaConsumerTest` to verify the new behavior. One minor compatibility implication worth mentioning is apparent from the change I made in `DynamicBrokerReconfigurationTest`. Previously it was possible to assume that all partitions had a fetch position after `poll()` completed with a non-empty assignment. This assumption is no longer generally true, but you can force the positions to be updated using the `position()` API which still blocks indefinitely until a position is available.
Note that this this patch also removes the logic to cache committed offsets in `SubscriptionState` since it was no longer needed (the consumer's `committed()` API always does an offset lookup anyway). In addition to avoiding the complexity of maintaining the cache, this avoids wasteful offset lookups to refresh the cache when `commitAsync()` is used.
Reviewers: Rajini Sivaram <rajinisivaram@googlemail.com>
* do not use static properties
* use new object to take appID
* capture timeout exception inside condition
Reviewers: Matthias J. Sax <matthias@confluent.io>
JsonConverter should use object equality rather than reference equality in `convertToJson`.
Reviewers: Bartlomiej Tartanus <bartektartanus@gmail.com>, Randall Hauch <rhauch@gmail.com>, Jason Gustafson <jason@confluent.io>
Corrected the parsing of invalid list values. A list can only be parsed if it contains elements that have a common type, and a map can only be parsed if it contains keys with a common type and values with a common type.
Reviewers: Arjun Satish <arjun@confluent.io>, Magesh Nandakumar <magesh.n.kumar@gmail.com>, Konstantine Karantasis <konstantine@confluent.io>, Jason Gustafson <jason@confluent.io>
* Rolling kill-restart Streams instances with brokers unavailable temporarily, and validate that the streams can still complete the restart process
Reviewers: Bill Bejeck <bill@confluent.io>, Matthias J. Sax <matthias@confluent.io>
Changed call to use the overload of ConnectSchema.validateValue() method with the field name passed in. Ensure that field in put call is not null.
Reviewers: Randall Hauch <rhauch@gmail.com>, Jason Gustafson <jason@confluent.io>
The commits for KIP-145 (KAFKA-5142) changed how the Connect workers instantiate and configure the Converters, and also added the ability to do the same for the new HeaderConverters. However, the last few commits removed the default value for the `converter.type` property for Converters and HeaderConverters, and this broke how the internal converters were being created.
This change corrects the behavior so that the `converter.type` property is always set by the worker (or by the Plugins class), which means the existing Converter implementations will not have to do this. The built-in JsonConverter, ByteArrayConverter, and StringConverter also implement HeaderConverter which implements Configurable, but the Worker and Plugins methods do not yet use the `Configurable.configure(Map)` method and instead still use the `Converter.configure(Map,boolean)`.
Several tests were modified, and a new PluginsTest was added to verify the new behavior in Plugins for instantiating and configuring the Converter and HeaderConverter instances.
Author: Randall Hauch <rhauch@gmail.com>
Reviewers: Konstantine Karantasis <konstantine@confluent.io>, Ewen Cheslack-Postava <ewen@confluent.io>
Closes#4512 from rhauch/kafka-6513
Author: Matthias J. Sax <matthias@confluent.io>
Reviewer: Damian Guy <damian@confluent.io>, Bill Bejeck <bill@confluent.io>, Guozhang Wang <guozhang@confluent.io>
If an exception is encountered while sending data to a client connection, that connection is disconnected. If there are staged receives for that connection, they are tracked to process those records. However, if the exception was encountered during processing a `RequestChannel.Request`, the `KafkaChannel` for that connection is muted and won't be processed.
Disable processing of outstanding staged receives if a send fails. This stops the leak of the memory for pending requests and the file descriptor of the TCP socket.
Test that a channel is closed when an exception is raised while writing to a socket that has been closed by the client. Since sending a response requires acks != 0, allow specifying the required acks for test requests in SocketServerTest.scala.
Author: Graham Campbell <graham.campbell@salesforce.com>
Reviewers: Jason Gustafson <jason@confluent.io>, Rajini Sivaram <rajinisivaram@googlemail.com>, Ismael Juma <ismael@juma.me.uk>, Ted Yu <yuzhihong@gmail.com>
1. Handle listener-not-found in MetadataCache since this can occur when listeners are being updated. To avoid breaking clients, this is handled in the same way as broker-not-available so that clients may retry.
2. Set retries=1000 for listener reconfiguration tests to avoid transient failures when metadata cache has not been updated
3. Remove IdlePercent metric when Processor is deleted, add test
4. Reduce log segment size used during reconfiguration to avoid timeout while waiting for log rolling
5.Test markPartitionsForTruncation after fetcher thread resize
6. Move per-processor ResponseQueueSize metric back to RequestChannel.
Reviewers: Ismael Juma <ismael@juma.me.uk>, Jason Gustafson <jason@confluent.io>
Adds web page docs for KIP-247
Author: Matthias J. Sax <matthias@confluent.io>
Reviewers: Guozhang Wang <guozhang@confluent.io>, Bill Bejeck <bill@confluent.io>, Joel Hamill <joel@confluent.io>, Damian Guy <damian@confluent.io>
Author: Matthias J. Sax <matthias@confluent.io>
Reviewers: Bill Bejeck <bill@confluent.io>, Joel Hamill <joel@confluent.io>, Guozhang Wang <guozhang@confluent.io>
Currently `maybeAutoCommitOffsetsAsync` does not try to find the coordinator if it is unknown. As a result, asynchronous auto-commits will fail indefinitely. This patch changes the behavior to add coordinator discovery to the async auto-commit path.
* MINOR: Update gradle, jackson and jacoco
- Gradle update adds support for Java 10
- Jacoco update adds support for Java 9
- Jackson bug fix update adds more serialization
robustness checks
* Update Jetty
Reviewers: Rajini Sivaram <rajinisivaram@googlemail.com>
Konstantine Karantasis
Jiangjie (Becket) Qin
Rajini Sivaram
ying-zheng
Yaswanth Kumar
Matthias J. Sax
Filipe Agapito
Attila Sasvari
dan norwood
Jason Gustafson
Eugene Sevastyanov
Manikumar Reddy O
Lucas Wang
Sebastian Bauersfeld
Robert Yokota
Bill Bejeck
Guozhang Wang
ConcurrencyPractitioner
Magnus Reftel
lisa2lisa
Benedict Jin
Randall Hauch
Kamal C
Michael Arndt
Joel Hamill
Dmitry Minkovsky
Vahid Hashemian
xin
Jeremy Custenborder
parafiend
Daniel Wojda
huxi
Ismael Juma