[
https://issues.apache.org/jira/browse/KAFKA-4213?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15517875#comment-15517875
]
ASF GitHub Bot commented on KAFKA-4213:
---------------------------------------
GitHub user apurvam opened a pull request:
https://github.com/apache/kafka/pull/1903
KAFKA-4213: First set of system tests for replication throttling
Added the first set of system tests for replication quotas. These tests
validate throttling behavior during partition reassigment.
Along with this patch are fixes to the test framework which include:
1. KakfaService.verify_replica_reassignment: this method was a no-op and
would always return success, as explained in KAFKA-4204. This patch adds a
workaround to the problems mentioned there, by grepping correctly for success,
failure, and 'in progress' states of partition reassignment.
2.ProduceConsumeValidateTest.annotate_missing_messages would call
missing.pop() to enumerate the first 20 missing messages. This meant that all
future counts of what is actually missing would be off by 20, leading to the
impression of data loss.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/apurvam/kafka throttling-tests
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/kafka/pull/1903.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #1903
----
commit 372db72c2da55dd2aba70019b258429855832804
Author: Apurva Mehta <[email protected]>
Date: 2016-09-09T18:53:42Z
Merge remote-tracking branch 'apache/trunk' into trunk
commit ae912d444d3fb63c2e5487f88949408e0b1207e9
Author: Jason Gustafson <[email protected]>
Date: 2016-09-09T20:44:55Z
KAFKA-3807; Fix transient test failure caused by race on future completion
Author: Jason Gustafson <[email protected]>
Reviewers: Dan Norwood <[email protected]>, Ismael Juma
<[email protected]>
Closes #1821 from hachikuji/KAFKA-3807
commit d0a86ffdec330f6e7213a370287a2d81bb93e2bc
Author: Vahid Hashemian <[email protected]>
Date: 2016-09-10T07:16:23Z
KAFKA-4145; Avoid redundant integration testing in ProducerSendTests
Author: Vahid Hashemian <[email protected]>
Reviewers: Ismael Juma <[email protected]>
Closes #1842 from vahidhashemian/KAFKA-4145
commit 42b5583561895e308063ed9e2186d83c83ca35d8
Author: Jason Gustafson <[email protected]>
Date: 2016-09-11T07:46:20Z
KAFKA-4147; Fix transient failure in
ConsumerCoordinatorTest.testAutoCommitDynamicAssignment
Author: Jason Gustafson <[email protected]>
Reviewers: Ismael Juma <[email protected]>
Closes #1841 from hachikuji/KAFKA-4147
commit e7697ad0ab0f292ad1e29d9a159d113574bfcf67
Author: Eric Wasserman <[email protected]>
Date: 2016-09-12T01:45:05Z
KAFKA-1981; Make log compaction point configurable
Now uses LogSegment.largestTimestamp to determine age of segment's messages.
Author: Eric Wasserman <[email protected]>
Reviewers: Jun Rao <[email protected]>
Closes #1794 from ewasserman/feat-1981
commit b36034eaa4eb284fafddb1a7507a2cf187993e62
Author: Damian Guy <[email protected]>
Date: 2016-09-12T04:00:32Z
MINOR: catch InvalidStateStoreException in QueryableStateIntegrationTest
A couple of the tests may transiently fail in QueryableStateIntegrationTest
as they are not catching InvalidStateStoreException. This exception is expected
during rebalance.
Author: Damian Guy <[email protected]>
Reviewers: Eno Thereska, Guozhang Wang
Closes #1840 from dguy/minor-fix
commit 642b709f919a02379f9d0c9313586b02d179ca78
Author: Tim Brooks <[email protected]>
Date: 2016-09-13T03:28:01Z
KAFKA-2311; Make KafkaConsumer's ensureNotClosed method thread-safe
Here is the patch on github ijuma.
Acquiring the consumer lock (the single thread access controls) requires
that the consumer be open. I changed the closed variable to be volatile so that
another thread's writes will visible to the reading thread.
Additionally, there was an additional check if the consumer was closed
after the lock was acquired. This check is no longer necessary.
This is my original work and I license it to the project under the
project's open source license.
Author: Tim Brooks <[email protected]>
Reviewers: Jason Gustafson <[email protected]>
Closes #1637 from tbrooks8/KAFKA-2311
commit ca539df5887bdfdbe86ba45f5514ed54b3b648d4
Author: Dong Lin <[email protected]>
Date: 2016-09-14T00:33:54Z
KAFKA-4158; Reset quota to default value if quota override is deleted
Author: Dong Lin <[email protected]>
Reviewers: Joel Koshy <[email protected]>, Jiangjie Qin
<[email protected]>
Closes #1851 from lindong28/KAFKA-4158
commit ba712d29eb2880fbf1709b5d0921028735a09f68
Author: Ismael Juma <[email protected]>
Date: 2016-09-14T16:16:29Z
MINOR: Give a name to the coordinator heartbeat thread
Followed the same naming pattern as the producer sender thread.
Author: Ismael Juma <[email protected]>
Reviewers: Jason Gustafson
Closes #1854 from ijuma/heartbeat-thread-name
commit 7c0f9b70e4e1ff643d953af50ca70e2d448ef431
Author: David Chen <[email protected]>
Date: 2016-09-14T17:38:40Z
KAFKA-4162: Fixed typo "rebalance"
Author: David Chen <[email protected]>
Reviewers: Ewen Cheslack-Postava <[email protected]>
Closes #1853 from mvj3/KAFKA-4162
commit 5ec0ffb32ef556ff22b24ad239f9aa546ac9783a
Author: Jason Gustafson <[email protected]>
Date: 2016-09-15T01:04:58Z
KAFKA-4172; Ensure fetch responses contain the requested partitions
Author: Jason Gustafson <[email protected]>
Reviewers: Ismael Juma <[email protected]>
Closes #1857 from hachikuji/KAFKA-4172
commit 8792ef05dc6253dfb1b673832cc0030e8bd3f075
Author: Jason Gustafson <[email protected]>
Date: 2016-09-15T05:31:52Z
KAFKA-4160: Ensure rebalance listener not called with coordinator lock
Author: Jason Gustafson <[email protected]>
Reviewers: Guozhang Wang <[email protected]>
Closes #1855 from hachikuji/KAFKA-4160
commit 57a69d9cab6884e0ed67f2a91320fc918034bfd2
Author: Damian Guy <[email protected]>
Date: 2016-09-15T15:57:48Z
HOTFIX: fix KafkaStreams SmokeTest
Set the NUM_STREAM_THREADS_CONFIG = 1 in SmokeTestClient as we get locking
issues when we have NUM_STREAM_THREADS_CONFIG > 1 and we have Standby Tasks,
i.e., replicas. This is because the Standby Tasks can be assigned to the same
KafkaStreams instance as the active task, hence the directory is locked
Author: Damian Guy <[email protected]>
Reviewers: Eno Thereska, Guozhang Wang
Closes #1861 from dguy/fix-smoketest
commit 5f555091bdd04cb49acf3da40ca1de84c731f6cb
Author: Bill Bejeck <[email protected]>
Date: 2016-09-16T00:08:00Z
KAFKA-4131; Multiple Regex KStream-Consumers cause Null pointer exception
Fix for bug outlined in KAFKA-4131
Author: bbejeck <[email protected]>
Reviewers: Damian Guy, Guozhang Wang
Closes #1843 from bbejeck/KAFKA-4131_mulitple_regex_consumers_cause_npe
commit bed93e182a52909a108849a3960123b592937653
Author: Ben Stopford <[email protected]>
Date: 2016-09-16T05:25:56Z
KAFKA-1464; Add a throttling option to the Kafka replication
This applies to Replication Quotas
based on KIP-73
[(link)](https://cwiki.apache.org/confluence/display/KAFKA/KIP-73+Replication+Quotas)
originally motivated by KAFKA-1464.
System Tests Run:
https://jenkins.confluent.io/job/system-test-kafka-branch-builder/544/
**This first PR demonstrates the approach**.
**_Overview of Change_**
The guts of this change are relatively small. Throttling occurs on both
leader and follower sides. A single class tracks the throttled throughput in
and out of each broker (**_ReplicationQuotaManager_**).
On the follower side, the Follower Throttled Rate is calculated as fetch
responses arrive. Then, before the next fetch request is sent, we check to see
if the quota is violated, removing throttled partitions from the request if it
is. This is all encapsulated in a few lines of code in the
**_ReplicaFetcherThread_**. There is existing code to handle temporal back off,
if the request ends up being empty.
On the leader side it's a little more complex. When a fetch request arrives
in the leader, it is built, partition by partition, in
**_ReplicaManager.readFromLocalLog_**. As we put each partition into the fetch
response, we check if the total size fits in the current quota. If the quota is
exceeded, the partition will not be added to the fetch response. Importantly,
we don't increase the quota at this point, we just check to see if the bytes
will fit.
Now, if there aren't enough bytes to send the response immediately, which
is common if we're catching up and throttled, then the request will be put in
purgatory. I've added some simple code to **_DelayedFetch_** to handle
throttled partitions (throttled partitions are checked against the quota,
rather than the messages available in the log).
When the delayed fetch completes, and exits purgatory,
_**ReplicaManager.readFromLocalLog**_ will be called again. This is why
_**ReplicaManager.readFromLocalLog**_ does not actually increase the quota, it
just checks whether enough bytes are available for a partition.
Finally, when there are enough bytes to be sent, or the delayed fetch times
out, the response will be sent. Before it is sent the throttled-outbound-rate
is increased, based on the size of throttled partitions being sent. This is at
the end of _**KafkaApis.handleFetchRequest**_, exactly where client quotas are
recorded.
There is an acceptance test which asserts the whole throttling process
stabilises on the desired value. This covers a number of use cases including
many-to-many replication. See **_ReplicationQuotaTest_**.
Note:
It should be noted that this protocol can over-request. The request is
built, based on the quota at time t1 (_ReplicaManager.readFromLocalLog_). The
bytes in the response are recorded at time t2 (end of
_KafkaApis.handleFetchRequest_), where t2 > t1. For this reason I originally
included an OverRequestedRate as a JMX metric, but testing has not seen
revealed any obvious issue. Over-requesting is quickly compensated by
subsequent requests, stabilising close to the quota value.
_**Main stuff left to do:**_
- The fetch size is currently unbounded. This will be addressed in KIP-74,
but we need to ensure this ensures requests don’t go beyond the throttle window.
- There are two failures showing up in the system tests on this branch:
StreamsSmokeTest.test_streams (which looks like it fails regularly) and
OffsetValidationTest.test_broker_rolling_bounce (which I need to look into)
_**Stuff left to do that could be deferred:**_
- Add the extra metrics specified in the KIP.
- There are no system tests.
- There is no validation for the cluster size / throttle combination that
could lead to ISR dropouts
Author: Ben Stopford <[email protected]>
Reviewers: Ismael Juma <[email protected]>, Apurva Mehta
<[email protected]>, Jun Rao <[email protected]>
Closes #1776 from benstopford/rep-quotas-v2
commit 7f3f0b1e511cc2579e8c99596b206e51feeb078a
Author: Damian Guy <[email protected]>
Date: 2016-09-16T16:58:36Z
KAFKA-3776: Unify store and downstream caching in streams
This is joint work between dguy and enothereska. The work implements
KIP-63. Overview of main changes:
- New byte-based cache that acts as a buffer for any persistent store and
for forwarding changes downstream.
- Forwarding record path changes: previously a record in a task completed
end-to-end. Now it may be buffered in a processor node while other records
complete in the task.
- Cleanup and state stores and decoupling of cache from state store and
forwarding.
- More than 80 new unit and integration tests.
Author: Damian Guy <[email protected]>
Author: Eno Thereska <[email protected]>
Reviewers: Matthias J. Sax, Guozhang Wang
Closes #1752 from enothereska/KAFKA-3776-poc
commit 5ea969a0d67c7df8344df819076baf6f54570cdb
Author: Randall Hauch <[email protected]>
Date: 2016-09-16T21:55:46Z
KAFKA-4183; Corrected Kafka Connect's JSON Converter to properly convert
from null to logical values
The `JsonConverter` class has `LogicalTypeConverter` implementations for
Date, Time, Timestamp, and Decimal, but these implementations fail when the
input literal value (deserialized from the message) is null.
Test cases were added to check for these cases, and these failed before the
`LogicalTypeConverter` implementations were fixed to consider whether the
schema has a default value or is optional, similarly to how the
`JsonToConnectTypeConverter` implementations do this. Once the fixes were made,
the new tests pass.
Author: Randall Hauch <[email protected]>
Reviewers: Shikhar Bhushan <[email protected]>, Jason Gustafson
<[email protected]>
Closes #1867 from rhauch/kafka-4183
commit 8b549e8f6485256cc586ba01ac5178871af21a65
Author: Shikhar Bhushan <[email protected]>
Date: 2016-09-16T22:54:33Z
KAFKA-4173; SchemaProjector should successfully project missing Struct
field when target field is optional
Author: Shikhar Bhushan <[email protected]>
Reviewers: Konstantine Karantasis <[email protected]>, Jason
Gustafson <[email protected]>
Closes #1865 from shikhar/kafka-4173
commit c9cec1bdc6abf0058319fae0a944790f3ad4fc7d
Author: Sumit Arrawatia <[email protected]>
Date: 2016-09-17T03:10:13Z
KAFKA-4093; Cluster Id (KIP-78)
This PR implements KIP-78:Cluster Identifiers
[(link)](https://cwiki.apache.org/confluence/display/KAFKA/KIP-78%3A+Cluster+Id#KIP-78:ClusterId-Overview)
and includes the following changes:
1. Changes to broker code
- generate cluster id and store it in Zookeeper
- update protocol to add cluster id to metadata request and response
- add ClusterResourceListener interface, ClusterResource class and
ClusterMetadataListeners utility class
- send ClusterResource events to the metric reporters
2. Changes to client code
- update Cluster and Metadata code to support cluster id
- update clients for sending ClusterResource events to interceptors,
(de)serializers and metric reporters
3. Integration tests for interceptors, (de)serializers and metric reporters
for clients and for protocol changes and metric reporters for broker.
4. System tests for upgrading from previous versions.
Author: Sumit Arrawatia <[email protected]>
Author: Ismael Juma <[email protected]>
Reviewers: Jun Rao <[email protected]>, Ismael Juma <[email protected]>
Closes #1830 from arrawatia/kip-78
commit 378e30c677c4f9d734ecd6858a925ff309c2ceb1
Author: Rajini Sivaram <[email protected]>
Date: 2016-09-17T17:06:05Z
KAFKA-3492; Secure quotas for authenticated users
Implementation and tests for secure quotas at <user> and <user, client-id>
levels as described in KIP-55. Also adds dynamic default quotas for
<client-id>, <user> and <user-client-id>. For each client connection, the most
specific quota matching the connection is used, with user quota taking
precedence over client-id quota.
Author: Rajini Sivaram <[email protected]>
Reviewers: Jun Rao <[email protected]>
Closes #1753 from rajinisivaram/KAFKA-3492
commit a78dabbc5a70e0be829d0143258c0c719ebfda74
Author: Eno Thereska <[email protected]>
Date: 2016-09-17T21:43:43Z
HOTFIX: Increase timeout for bounce test
Author: Eno Thereska <[email protected]>
Reviewers: Ismael Juma <[email protected]>
Closes #1874 from enothereska/hotfix-bounce-test
commit ce6f8a6ef73a54989e95d6ad6e4022e192392a8c
Author: Matthias J. Sax <[email protected]>
Date: 2016-09-17T21:45:29Z
HOTFIX: changed quickstart donwload from 0.10.0.0 to 0.10.0.1
Author: Matthias J. Sax <[email protected]>
Reviewers: Ismael Juma <[email protected]>
Closes #1869 from mjsax/hotfix-doc
commit 751622eceaf4eed5532b1e44f1738cbd81ae359d
Author: Grant Henke <[email protected]>
Date: 2016-09-17T21:47:56Z
KAFKA-4157; Transient system test failure in
replica_verification_test.test_replica_lags
…t.test_replica_lags
Author: Grant Henke <[email protected]>
Reviewers: Ashish Singh <[email protected]>, Ismael Juma
<[email protected]>
Closes #1849 from granthenke/replica-verification-fix
commit 639b7cd133d62275f6ada3344579ca7f755ef2fb
Author: Jaikiran Pai <[email protected]>
Date: 2016-09-17T22:01:32Z
MINOR: Update the README.md to include a note about GRADLE_USER_HOME
Trying to build the source and publish it to internal Maven repo, I ran
into an issue that I explain in the mailing list discussion here
https://www.mail-archive.com/devkafka.apache.org/msg56359.html.
The commit here updates the README.md to make a note that the
GRADLE_USER_HOME environment variable plays a role in deciding which file to
add the maven configs to.
Author: Jaikiran Pai <[email protected]>
Reviewers: Ismael Juma <[email protected]>
Closes #1837 from jaikiran/readme-update-grade-user-home
commit 6965270a84eee075adf466bd949ae7f1ff41e579
Author: Andrey Neporada <[email protected]>
Date: 2016-09-18T16:12:53Z
KAFKA-2063; Bound fetch response size (KIP-74)
This PR is implementation of
[KIP-74](https://cwiki.apache.org/confluence/display/KAFKA/KIP-74%3A+Add+Fetch+Response+Size+Limit+in+Bytes)
which is originally motivated by
[KAFKA-2063](https://issues.apache.org/jira/browse/KAFKA-2063).
Author: Andrey Neporada <[email protected]>
Author: Ismael Juma <[email protected]>
Reviewers: Jun Rao <[email protected]>, Jiangjie Qin <[email protected]>,
Jason Gustafson <[email protected]>
Closes #1812 from nepal/kip-74
commit bdc95ea18c01374d26f77eaa157c58f8f30e6edf
Author: Luke Zaparaniuk <[email protected]>
Date: 2016-09-18T21:59:00Z
MINOR: Fix reference to argument in `LogSegment.translateOffset`
Changed the lowerBound argument reference in the summary comment of the
translateOffset method to match the actual argument name: startingFilePosition.
Author: Luke Zaparaniuk <[email protected]>
Reviewers: Ismael Juma <[email protected]>
Closes #1876 from lukezaparaniuk/patch-1
commit 2a23e81f0d14d55684f6df4d3a717affcb40ab52
Author: Damian Guy <[email protected]>
Date: 2016-09-19T17:28:58Z
HOTFIX: logic in QuerybaleStateIntegrationTest.shouldBeAbleToQueryState
incorrect
The logic in `verifyCanGetByKey` was incorrect. It was
```
windowState.size() < keys.length &&
countState.size() < keys.length &&
System.currentTimeMillis() < timeout
```
but should be:
```
(windowState.size() < keys.length || countState.size() < keys.length) &&
System.currentTimeMillis() < timeout
```
Author: Damian Guy <[email protected]>
Reviewers: Guozhang Wang <[email protected]>
Closes #1879 from dguy/minor-fix-test
commit 21ab564f4dd8d3d7a443d74f444cee3b6eecc664
Author: Damian Guy <[email protected]>
Date: 2016-09-19T17:30:58Z
KAFKA-4175: Can't have StandbyTasks in KafkaStreams where
NUM_STREAM_THREADS_CONFIG > 1
standby tasks should be assigned per consumer not per process
Author: Damian Guy <[email protected]>
Reviewers: Eno Thereska, Guozhang Wang
Closes #1862 from dguy/kafka-4175
commit 184c0a6c04784491055033c747c13633df4ba1c0
Author: Damian Guy <[email protected]>
Date: 2016-09-19T18:00:53Z
KAFKA-4163: NPE in StreamsMetadataState during re-balance operations
During rebalance operations the Cluster object gets set to Cluster.empty().
This can result in NPEs when doing certain operation on StreamsMetadataState.
This should throw a StreamsException if the Cluster is empty as it is not yet
(re-)initialized
Author: Damian Guy <[email protected]>
Reviewers: Eno Thereska, Guozhang Wang
Closes #1845 from dguy/streams-meta-hotfix
commit 3e983af14c138c555a0443c94c3d05631073c2fd
Author: Rajini Sivaram <[email protected]>
Date: 2016-09-19T19:16:45Z
KAFKA-4079; Documentation for secure quotas
Details in KIP-55.
Author: Rajini Sivaram <[email protected]>
Reviewers: Jun Rao <[email protected]>
Closes #1847 from rajinisivaram/KAFKA-4079
----
> Add system tests for replication throttling (KIP-73)
> ----------------------------------------------------
>
> Key: KAFKA-4213
> URL: https://issues.apache.org/jira/browse/KAFKA-4213
> Project: Kafka
> Issue Type: Test
> Reporter: Apurva Mehta
> Assignee: Apurva Mehta
>
> Add system tests for replication throttling. The two main things to test are:
> 1. kafka-reassign-partitions: in this use case, a new broker is added to a
> cluster, and we are testing throttling of the partitions being replicated to
> this cluster. The '--throttle' option of the reassign partitions tool is what
> we want to test. we will invoke the tool with this option, and assert that
> the replication takes a minimum amount of time, based on the throttle and the
> amount of data being replicated.
> 2. kafka-configs: in this use case, we lost a broker of an existing cluster
> for whatever reason, and want to re-replicate data to it from some point in
> time. We want this re-replicated data to be throttled. Again, we will check
> that the re-replication took at least a certain amount of time based on the
> value of the throttle and the amount of data being replicated.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
