ddd652c6 standardized on "isolated" as the name for all the isolated
modes, and renamed kafkatest.services.kafka.quorum.remote_kraft to
isolated_kraft. However, the tests using remote_kraft weren't
updated, and are broken as a result. This is a simple search and
replace to fix those.
Reviewers: Colin P. McCabe <cmccabe@apache.org>
Standardize KRaft thread names.
- Always use kebab case. That is, "my-thread-name".
- Thread prefixes are just strings, not Option[String] or Optional<String>.
If you don't want a prefix, use the empty string.
- Thread prefixes end in a dash (except the empty prefix). Then you can
calculate thread names as $prefix + "my-thread-name"
- Broker-only components get "broker-$id-" as a thread name prefix. For example, "broker-1-"
- Controller-only components get "controller-$id-" as a thread name prefix. For example, "controller-1-"
- Shared components get "kafka-$id-" as a thread name prefix. For example, "kafka-0-"
- Always pass a prefix to KafkaEventQueue, so that threads have names like
"broker-0-metadata-loader-event-handler" rather than "event-handler". Prior to this PR, we had
several threads just named "EventHandler" which was not helpful for debugging.
- QuorumController thread name is "quorum-controller-123-event-handler"
- Don't set a thread prefix for replication threads started by ReplicaManager. They run only on the
broker, and already include the broker ID.
Standardize KRaft slf4j log prefixes.
- Names should be of the form "[ComponentName id=$id] ". So for a ControllerServer with ID 123, we
will have "[ControllerServer id=123] "
- For the QuorumController class, use the prefix "[QuorumController id=$id] " rather than
"[Controller <nodeId] ", to make it clearer that this is a KRaft controller.
- In BrokerLifecycleManager, add isZkBroker=true to the log prefix for the migration case.
Standardize KRaft terminology.
- All synonyms of combined mode (colocated, coresident, etc.) should be replaced by "combined"
- All synonyms of isolated mode (remote, non-colocated, distributed, etc.) should be replaced by
"isolated".
This patch removes test_kafka_version.py, which contains two tests at the moment. The first test verifies we can start a 0.8.2 cluster. The second verifies we can start a cluster with one node on 0.8.2 and another on the latest. These test are covered in greater depth by upgrade_test.py and downgrade_test.py.
Reviewers: José Armando García Sancio <jsancio@users.noreply.github.com>
This patch adds a sanity-check bounce system test for the case where we have 3
co-located KRaft controllers and fixes the system test code so that this case
will pass by starting brokers in parallel by default instead of serially. We
now also send SIGKILL to any running KRaft broker or controller nodes for the
co-located case when a majority of co-located controllers have been stopped --
otherwise they do not shutdown, and we spin for the 60 second timeout. Finally,
this patch adds the ability to specify that certain brokers should not be
started when starting the cluster, and then we can start those nodes at a later
time via the add_broker() method call; this is going to be helpful for KRaft
snapshot system testing.
We were not testing the 3 co-located KRaft controller case previously, and it
would not pass because the first Kafka node would never be considered started.
We were starting the Kafka nodes serially, and we decide that a node has
successfully started when it logs a particular message. This message is not
logged until the broker has identified the controller (i.e. the leader of the
KRaft quorum). There cannot be a leader until a majority of the KRaft quorum
has started, so with 3 co-located controllers the first node could never be
considered "started" by the system test.
Reviewers: Colin P. McCabe <cmccabe@apache.org>
Update the ZooKeeper version to v3.6.3. This requires adding dropwizard
as a new dependency.
Also, add Kafka v2.8.0 to the ducktape system test image.
Reviewers: Luke Chen <showuon@gmail.com>, Colin P. McCabe <cmccabe@apache.org>, Ismael Juma <ismael@juma.me.uk>
This patch adds support for running the ZooKeeper-based
kafka.security.authorizer.AclAuthorizer with KRaft clusters. Set the
authorizer.class.name config as well as the zookeeper.connect config while also
setting the typical KRaft configs (node.id, process.roles, etc.), and the
cluster will use KRaft for metadata and ZooKeeper for ACL storage. A system
test that exercises the authorizer is included.
This patch also changes "Raft" to "KRaft" in several system test files. It also
fixes a bug where system test admin clients were unable to connect to a cluster
with broker credentials via the SSL security protocol when the broker was using
that for inter-broker communication and SASL for client communication.
Reviewers: Colin P. McCabe <cmccabe@apache.org>, Ismael Juma <ismael@juma.me.uk>
Fix some cases where we were erroneously using the configuration of the inter broker
listener instead of the controller listener. Add the sasl.mechanism.controller.protocol
configuration key specified by KIP-631. Add some ducktape tests.
Reviewers: Colin P. McCabe <cmccabe@apache.org>, David Arthur <mumrah@gmail.com>, Boyang Chen <boyang@confluent.io>
Add the necessary test annotations to test the new KIP-500 quorum broker mode
in many of our ducktape tests. This mode is tested in addition to the classic
Apache ZooKeeper mode.
This PR also adds a new sanity_checks/bounce_test.py system test that runs
through a simple produce/bounce/produce series of events.
Finally, this PR adds @cluster annotations to dozens of system tests that were
missing them. The lack of this annotation was causing these tests to grab the
entire cluster of nodes. Adding the @cluster annotation dramatically reduced
the time needed to run these tests.
Reviewers: Colin P. McCabe <cmccabe@apache.org>, Ismael Juma <ismael@juma.me.uk>
Increase the amount of time available to the `test_verifiable_producer` test to login and get the process name for the verifiable producer from 5 seconds to 10 seconds.
We were seeing some test failures due to the assertion failing because the verifiable producer would complete before we could login, list the processes, and parse out the producer version. Previously, we were giving this operation 5 seconds to run, this PR bumps it up to 10 seconds.
I verified locally that this does not flake, but even at 5 seconds I wasn't seeing any flakes. Ultimately we should find a better strategy than racing to query the producer process (as outlined in the existing comments).
Reviewers: Jason Gustafson <jason@confluent.io>
For now, Kafka system tests use python2 which is outdated and not supported.
This PR upgrades python to the third version.
Reviewers: Ivan Daschinskiy, Mickael Maison <mickael.maison@gmail.com>, Magnus Edenhill <magnus@edenhill.se>, Guozhang Wang <wangguoz@gmail.com>
For consumers using SSL, this timeout includes the time to create and copy keystores and truststores and sometime it takes longer than 10s to complete the security setup before starting the consumer process.
Reviewers: Ismael Juma <ismael@juma.me.uk>
- Removed Scala consumers (`SimpleConsumer` and `ZooKeeperConsumerConnector`)
and their tests.
- Removed Scala request/response/message classes.
- Removed any mention of new consumer or new producer in the code
with the exception of MirrorMaker where the new.consumer option was
never deprecated so we have to keep it for now. The non-code
documentation has not been updated either, that will be done
separately.
- Removed a number of tools that only made sense in the context
of the Scala consumers (see upgrade notes).
- Updated some tools that worked with both Scala and Java consumers
so that they only support the latter (see upgrade notes).
- Removed `BaseConsumer` and related classes apart from `BaseRecord`
which is used in `MirrorMakerMessageHandler`. The latter is a pluggable
interface so effectively public API.
- Removed `ZkUtils` methods that were only used by the old consumers.
- Removed `ZkUtils.registerBroker` and `ZKCheckedEphemeral` since
the broker now uses the methods in `KafkaZkClient` and no-one else
should be using that method.
- Updated system tests so that they don't use the Scala consumers except
for multi-version tests.
- Updated LogDirFailureTest so that the consumer offsets topic would
continue to be available after all the failures. This was necessary for it
to work with the Java consumer.
- Some multi-version system tests had not been updated to include
recently released Kafka versions, fixed it.
- Updated findBugs and checkstyle configs not to refer to deleted
classes and packages.
Reviewers: Dong Lin <lindong28@gmail.com>, Manikumar Reddy <manikumar.reddy@gmail.com>
Ensures Kafka chroot exists in ZK when starting KafkaService so commands that use ZK and are executed before the first Kafka broker starts do not fail due to the missing chroot.
Also uses chroot with one test that also has security parameterizations so Kafka's test suite exercises these combinations. Previously no tests were exercising chroots.
Changes were validated using sanity_checks which include the chroot-ed test as well as some non-chroot-ed tests.
## Fixes race condition in TestVerifiableProducer sanity test:
The test starts a producer, waits for at least 5 acks, and then
logs in to the worker to grep for the producer process to figure
out what version it is running.
The problem was that the producer was set up to produce 1000 messages
at a rate of 1000 msgs/s and then exit. This means it will have a
typical runtime slightly above 1 second.
Logging in to the vagrant instance might take longer than that thus
resulting in the process grep to fail, failing the test.
This commit doesn't really fix the issue - a proper fix would be to tell
the producer to stick around until explicitly killed - but it increases
the chances of the test passing, at the expense of a slightly longer
runtime.
## Improves error reporting when is_version() fails
Author: Magnus Edenhill <magnus@edenhill.se>
Reviewers: Apurva Mehta <apurva@confluent.io>, Ewen Cheslack-Postava <ewen@confluent.io>
Closes#2765 from edenhill/trunk
…0.x and not 0.8
Author: Colin P. Mccabe <cmccabe@confluent.io>
Reviewers: Ewen Cheslack-Postava <ewen@confluent.io>
Closes#2602 from cmccabe/KAFKA-4809
Runs sanity test and one replication test using SASL/SCRAM.
Author: Rajini Sivaram <rajinisivaram@googlemail.com>
Reviewers: Ewen Cheslack-Postava <ewen@confluent.io>, Ismael Juma <ismael@juma.me.uk>
Closes#2355 from rajinisivaram/KAFKA-4590
Updates to take advantage of soon-to-be-released ducktape features.
Author: Geoff Anderson <geoff@confluent.io>
Author: Ewen Cheslack-Postava <me@ewencp.org>
Reviewers: Ewen Cheslack-Postava <ewen@confluent.io>
Closes#1834 from granders/systest-parallel-friendly
Update system test method signatures and method calls to use the new consumer by default.
Author: Vahid Hashemian <vahidhashemian@us.ibm.com>
Reviewers: Jason Gustafson <jason@confluent.io>
Closes#2060 from vahidhashemian/KAFKA-4211
This patch adds logic for the following:
- remove hard-coded paths to various scripts and jars in kafkatest service classes
- provide a mechanism for overriding path resolution logic with a "pluggable" path resolver class
Author: Geoff Anderson <geoff@confluent.io>
Reviewers: Ewen Cheslack-Postava <ewen@confluent.io>
Closes#1245 from granders/configurable-install-path
This actually removes joins altogether, as well as references to self.worker_threads, which is best left as an implementation detail in BackgroundThreadService.
This makes use of hachikuji 's recent ducktape patch, and updates ducktape dependency to 0.5.0.
Author: Geoff Anderson <geoff@confluent.io>
Reviewers: Jason Gustafson <jason@confluent.io>, Ewen Cheslack-Postava <ewen@confluent.io>
Closes#1297 from granders/KAFKA-3581-systest-add-join-timeout
Run a sanity test with SASL/PLAIN and a couple of replication tests with SASL/PLAIN and multiple mechanisms.
Author: Rajini Sivaram <rajinisivaram@googlemail.com>
Reviewers: Ismael Juma <ismael@juma.me.uk>, Ewen Cheslack-Postava <ewen@confluent.io>
Closes#1282 from rajinisivaram/KAFKA-2693
Author: Ismael Juma <ismael@juma.me.uk>
Author: Geoff Anderson <geoff@confluent.io>
Reviewers: Geoff Anderson <geoff@confluent.io>, Ewen Cheslack-Postava <ewen@confluent.io>
Closes#1173 from ijuma/kafka-3490-multiple-version-support-perf-tests
Three main tests:
1. Setup: Producer (0.8) → Kafka Cluster → Consumer (0.8)
First rolling bounce: Set inter.broker.protocol.version = 0.8 and message.format.version = 0.8
Second rolling bonus, use latest (default) inter.broker.protocol.version and message.format.version
2. Setup: Producer (0.9) → Kafka Cluster → Consumer (0.9)
First rolling bounce: Set inter.broker.protocol.version = 0.9 and message.format.version = 0.9
Second rolling bonus, use latest (default) inter.broker.protocol.version and message.format.version
3. Setup: Producer (0.9) → Kafka Cluster → Consumer (0.9)
First rolling bounce: Set inter.broker.protocol.version = 0.9 and message.format.version = 0.9
Second rolling bonus: use inter.broker.protocol.version = 0.10 and message.format.version = 0.9
Plus couple of variations of these tests using old/new consumer and no compression / snappy compression.
Author: Anna Povzner <anna@confluent.io>
Reviewers: Ewen Cheslack-Postava <ewen@confluent.io>
Closes#980 from apovzner/kafka-3201-02
This PR adds failover to simple end to end mirror maker test
Marked as WIP for 2 reasons:
- We may want to add a couple more test cases where kafka is being used to store offsets
- There appears to be a test failure in the hard failover case
Author: Geoff Anderson <geoff@confluent.io>
Reviewers: Ewen Cheslack-Postava
Closes#427 from granders/KAFKA-2258-mirrormaker-test
Run sanity check, replication tests and benchmarks with SASL/Kerberos using MiniKdc.
Author: Rajini Sivaram <rajinisivaram@googlemail.com>
Reviewers: Geoff Anderson <geoff@confluent.io>, Jun Rao <junrao@gmail.com>
Closes#358 from rajinisivaram/KAFKA-2644
ewencp gwenshap
This needs some refactoring to avoid the duplicated code between replication test and upgrade test, but in shape for initial feedback.
I'm interested in feedback on the added `KafkaConfig` class and `kafka_props` file. This addition makes it:
- easier to attach different configs to different nodes (e.g. during broker upgrade process)
- easier to reason about the configuration of a particular node
Notes:
- in the default values in the KafkaConfig class, I removed many properties which were in kafka.properties before. This is because most of those properties were set to what is already the default value.
- when running non-trunk VerifiableProducer, I append the trunk tools jar to the classpath, and run it with the non-trunk kafka-run-class.sh script
Author: Geoff Anderson <geoff@confluent.io>
Reviewers: Dong Lin, Ewen Cheslack-Postava
Closes#229 from granders/KAFKA-1888-upgrade-test
Parametrize console consumer sanity test, replication tests and benchmarks tests to run with both PLAINTEXT and SSL.
Author: Rajini Sivaram <rajinisivaram@googlemail.com>
Reviewers: Geoff Anderson, Ewen Cheslack-Postava, Guozhang Wang
Closes#271 from rajinisivaram/KAFKA-2581
Tests standalone mode by running separate source and sink connectors, catting
data into the source file, and validating the output in the sink file. Restarts
the service to verify that clean restarts will result in tasks resuming where
they left off.
Author: Ewen Cheslack-Postava <me@ewencp.org>
Reviewers: Geoff Andreson, Gwen Shapira
Closes#150 from ewencp/kafka-2377-copycat-system-test
Added MirrorMaker service and a few corresponding sanity checks, as well as necessary config template files. A few additional updates to accomodate the change in wait_until from ducktape0.2.0->0.3.0
Author: Geoff Anderson <geoff@confluent.io>
Reviewers: Ewen Cheslack-Postava, Gwen Shapira
Closes#148 from granders/KAFKA-2439 and squashes the following commits:
c7c3ebd [Geoff Anderson] MirrorMaker now can run as multi-node service. Added kill -9 to various clean_node methods.
1e806f2 [Geoff Anderson] Various cleanups per review.
1b4b049 [Geoff Anderson] Added MirrorMaker service and a few corresponding sanity checks, as well as necessary config template files. A few additional updates to accomodate the change in wait_until from ducktape0.2.0->0.3.0
console consumer writes to System.out, while (some) log4j loggers operate in other threads.
This occasionally led to funky interleaved output which disrupted parsing of consumed messages by ConsoleConsumerService, leading to spurious test failures.
This fix directs log output to a separate file.
Author: Geoff Anderson <geoff@confluent.io>
Reviewers: Ewen Cheslack-Postava
Closes#123 from granders/KAFKA-2408 and squashes the following commits:
247b0e0 [Geoff Anderson] Updated line counting to use wc -l
66d6f4f [Geoff Anderson] lower -> uperrcase constants
e67f554 [Geoff Anderson] Changed incorrect license header
af67e01 [Geoff Anderson] Merged in upstream trunk
8f89044 [Geoff Anderson] Added another lifecycle check. Wait for log file to exist before exmaning contents.
521a84b [Geoff Anderson] Updated console consumer to directo log output directly to file rather than stdout
David Jacot
PoAn Yang
TengYao Chi
Yung
Greg Harris
Shay Elkin
Colin Patrick McCabe
Jason Gustafson
Cheng Tan
Ron Dagostino
Gardner Vickers
Nikolay
Rajini Sivaram
Ismael Juma
Ewen Cheslack-Postava
Magnus Edenhill
Colin P. Mccabe
Geoff Anderson
Vahid Hashemian
Anna Povzner
Ashish Singh