Junkai Xue [Fri, 15 Apr 2022 04:44:22 +0000 (21:44 -0700)]
[maven-release-plugin] prepare for next development iteration
Junkai Xue [Fri, 15 Apr 2022 04:44:07 +0000 (21:44 -0700)]
[maven-release-plugin] prepare release helix-1.0.3
Junkai Xue [Fri, 15 Apr 2022 04:19:01 +0000 (21:19 -0700)]
Change apache source url
Junkai Xue [Fri, 15 Apr 2022 03:27:07 +0000 (20:27 -0700)]
enable helix-front
Molly Gao [Wed, 6 Apr 2022 17:53:45 +0000 (10:53 -0700)]
Implement DefaultCloudEventCallbackImpl (#1995)
Implement a default callback implementation for Helix cloud event listeners.
Molly Gao [Wed, 6 Apr 2022 17:53:22 +0000 (10:53 -0700)]
Fix TestCloudEventCallbackProperty by bypassing connecting to zk (#2017)
Due to logical change in ZKHelixManager constructor in a #1986, TestCloudEventCallbackProperty broke because in this test it doesn't connect to zookeeper server.
To fix this test, we separated MockCloudEventAwareHelixManager (previously called MockEventAwareZKHelixManager, nested inside TestCloudEventCallbackProperty)into a class, and include all and only the logics related to cloud events in MockCloudEventAwareHelixManager. More detailed, we mock a cloud config object retrieved from zk to bypass connection to zk.
Molly Gao [Mon, 21 Mar 2022 20:31:35 +0000 (13:31 -0700)]
Enable HelixManager as an event listener (#1978)
Make helix manager cloud event aware by registering a cloud event listener when connect Helix manager
Molly Gao [Thu, 10 Mar 2022 22:05:33 +0000 (14:05 -0800)]
Add event handler and event listener interface (#1976)
This commit creates a skeleton for event handling framework and adds the following classes:
CloudEventListener interface
CloudEventHandler class
CloudEventHandlerFactory class
Henri Hagberg [Thu, 7 Apr 2022 14:09:49 +0000 (17:09 +0300)]
Improve ZkClientMonitor and ZkClientPathMonitor performance (#2021)
Previously, regex matches were used, which was inefficient. This commit does this following:
Replace String#matches with more efficient String#contains in ZkClientPathMonitor
Refactor record* methods in ZkClientMonitor to avoid repetition and simplify matching logic
xyuanlu [Wed, 6 Apr 2022 01:34:49 +0000 (18:34 -0700)]
add type and reason to cluster config (#2006)
Add type and reason for batch disable/enable instance
xyuanlu [Tue, 5 Apr 2022 21:01:02 +0000 (14:01 -0700)]
Refactor config string processing logic into a util class (#2015)
Refactor config string processing logic into a util class
Qi (Quincy) Qu [Tue, 5 Apr 2022 17:41:44 +0000 (10:41 -0700)]
Fix TestDropResourceMetricsReset (#2011)
Add a temp workaround by manually triggering a CurrentStageChange event for ExternalViewStage computation so that resource monitor can be cleaned up in unit test.
Molly Gao [Tue, 5 Apr 2022 17:13:01 +0000 (10:13 -0700)]
Populate helix cloud property using cloud config (#2005)
Currently when instantiating a zk helix manager, we retrieve cloud config from zk and replace entire HelixCloudProperty object, which may cause some fields that user pass in in HelixCloudProperty that are not included in cloud config get missing. This commit changes the logic to only populate fields in HelixCloudProperty with values that are present in cloud config, and leave other fields unchanged.
Junkai Xue [Mon, 4 Apr 2022 17:17:49 +0000 (10:17 -0700)]
Fix web build and deployment Issue (#2014)
xyuanlu [Mon, 4 Apr 2022 17:14:03 +0000 (10:14 -0700)]
fix Broken Logic in perPartitionHealthCheck (#2012)
fix Broken Logic in perPartitionHealthCheck
Hervé Boutemy [Sun, 3 Apr 2022 21:18:07 +0000 (23:18 +0200)]
fix repoducible builds issue (#2013)
Junkai Xue [Sat, 2 Apr 2022 23:29:28 +0000 (16:29 -0700)]
Update 0.9.9 to be replaced by 0.9.10
xyuanlu [Thu, 31 Mar 2022 23:19:00 +0000 (16:19 -0700)]
add new error message for customized partition check host connection error (#1984)
Add new error message for customized partition check host connection error
xyuanlu [Thu, 31 Mar 2022 23:18:25 +0000 (16:18 -0700)]
auto Exit MM for auto EMM test (#2003)
xyuanlu [Thu, 31 Mar 2022 23:18:11 +0000 (16:18 -0700)]
Add instance disable reason (#1993) (#2004)
Add instance disable reason
Qi (Quincy) Qu [Thu, 24 Mar 2022 16:51:46 +0000 (09:51 -0700)]
Implement deactivate rest API (#1988)
Implement deactivate REST API
Added a new REST API for deactivating cluster from supercluster.
xyuanlu [Thu, 24 Mar 2022 16:30:00 +0000 (09:30 -0700)]
Read cloud config from zk and propagate to HelixManagerProperty in ZkHelixManager constructor (#1986)
Read cloud config from zk and propagate to HelixManagerProperty in ZkHelixManager constructor.
Neal Sun [Mon, 14 Mar 2022 23:30:43 +0000 (16:30 -0700)]
Add Authorization Components to helix-rest (#1967) (#1981)
* Add Authorization Components to helix-rest
* Address some comments
Hunter Lee [Sat, 12 Mar 2022 23:22:21 +0000 (18:22 -0500)]
Use ZooKeeper 3.5.9 in zookeeper-api instead (#1977)
With the upgrade of apache zookeeper version, snappy-java was missing in the dependency. This commit adds snappy-java and removes unused imports in osgi declaration. Also, using ZooKeeper 3.5.9 in zookeeper-api instead because 3.6.0+ causes some tests to fail in zookeeper-api.
Ramin Bashizade [Wed, 9 Mar 2022 17:29:31 +0000 (09:29 -0800)]
Add 3 Zookeeper CreateMode types to AccessOption (#1975)
This commit adds the 3 missing CreateMode types to AccessOption
class in helix-core: CONTAINER, PERSISTENT_WITH_TTL, and
PERSISTENT_SEQUENTIAL_WITH_TTL.
rahulrane50 [Tue, 8 Mar 2022 21:42:44 +0000 (13:42 -0800)]
Upgrading Zookeeper version to 3.6.13 to enable zk client SSL/TLS
Upgrading Zookeeper version to 3.6.13 to enable zk client SSL/TLS support
Micah Stubbs [Thu, 3 Mar 2022 22:08:34 +0000 (14:08 -0800)]
[HELIX-862] s/maintainence/maintenance docs fix (#1968)
Qi (Quincy) Qu [Wed, 16 Feb 2022 17:55:56 +0000 (12:55 -0500)]
Add rest endpoint for virtual topology group (#1958)
Qi (Quincy) Qu [Tue, 8 Feb 2022 21:53:40 +0000 (16:53 -0500)]
Implement java API and utils for virtual topology group (#1935)
Add comment to VirtualTopologyGroupService.
Qi (Quincy) Qu [Thu, 3 Feb 2022 20:18:46 +0000 (12:18 -0800)]
Introduce VirtualTopologyGroup and its assignment logic with benchmark. (#1948)
* Cleanup unused assignment schemes and minor change.
* Further refactor and code cleanup.
Qi (Quincy) Qu [Sat, 29 Jan 2022 00:14:09 +0000 (16:14 -0800)]
Fix #1946 -- Refactor and move ClusterTopologyConfig
Move ClusterTopologyConfig from nested to a standalone class in helix/model and to be used by virtual topology group logic.
xyuanlu [Wed, 16 Feb 2022 21:52:48 +0000 (13:52 -0800)]
Use final remaining capacity when computing weighted score (#1961)
WAGED improvement: Use final remaining capacity when computing weighted score
xyuanlu [Tue, 15 Feb 2022 00:58:05 +0000 (16:58 -0800)]
Remove WAGED sorting for each assignment (#1959)
Improve WAGED sorting from n^2 to n*log(n)
xyuanlu [Mon, 7 Feb 2022 23:07:16 +0000 (15:07 -0800)]
remove log before write error message to ZNode (#1955)
Remove message logging before writing ZNode.
Komal Desai [Mon, 7 Feb 2022 23:06:57 +0000 (15:06 -0800)]
Fixes #1802 - messages intended for instances that are no longer in the cluster (#1951)
In MessageGenerationPhase.java, - process() method populates the list of live instances from cache.
But while generateMessage() method has the sessionIdMap information, it still goes through partition/resource/instance map without checking if instance is still part of the cluster or not.
It is possible that cache has stale entry but that logic needs to be worked separately. But while generating message, we should check if the instance is still there.
So this is a simple change. We need to still look further if cache is getting invalidated properly.
To make sure that the cache properly is handled/refreshed under instance being replaced or deletion - have filled another bug: #1956
Henri Hagberg [Thu, 3 Feb 2022 21:35:05 +0000 (23:35 +0200)]
Let logging framework format exception stack traces (#1954)
Where possible, logging calls are changed so that logging framework handles exception formatting instead of stack trace being manually formatted using Throwable#getStackTrace
Jiajun Wang [Wed, 2 Feb 2022 20:28:14 +0000 (12:28 -0800)]
Add new metrics to record ZNRecord compression count. (#1943)
This PR determines if a ZK write request is compressed by calling GZipCompressionUtil. This is an indirect method and can be inaccurate. So the decision is based on trade-offs.
Alternatively, the ZkClientMonitor can be passed into the serializer class and then report compressed write internally. However, this will require multiple changes in the serializer interfaces.
Due to the multiple layers (PathBasedZkSerializer, ZkSerializer) of serializer interfaces definition, it would be very costly to implement the alternative without major refactoring.
desaikomal [Mon, 31 Jan 2022 02:08:42 +0000 (18:08 -0800)]
Fix for - Stale message redundant logs
Avoid printing redundant log messages for unrelated partitions and resources.
desaikomal [Sat, 29 Jan 2022 00:16:16 +0000 (16:16 -0800)]
Fix Issue#1941 - Incorrect condition caused not to log error message
Properly populate the error log messages for partitions and resource names whose replica status is in ERROR state.
CVEDetect [Wed, 19 Jan 2022 15:17:11 +0000 (23:17 +0800)]
Fix CVE dependency issue (#1927)
Andrzej Hołowko [Wed, 19 Jan 2022 15:15:31 +0000 (16:15 +0100)]
Remove dependency to an old Jackson v1 library (org.codehaus.jackson:jackson-mapper-asl) affected by the critical vulnerability: CVE-2019-17267 (#1934)
Jiajun Wang [Wed, 19 Jan 2022 01:35:45 +0000 (17:35 -0800)]
Fix race condition in scheduler message processing logic. (#1930)
This PR aims to fix the race condition that happens during processing scheduler messages. The previous logic which dynamically delete task partitions in the scheduler message IdealState may cause conflicts and results in inconsistent message status update. Since updating the task partitions is not a necessary step, this PR removes the corresponding logic and simplify the message handling procedure.
This PR will help to stablize TestSchedulerMessage.java.
Henri Hagberg [Tue, 18 Jan 2022 22:30:13 +0000 (00:30 +0200)]
Daemonize ZkBucketDataAccessor GC_THREAD (#1936)
GC_THREAD (which is actually an ExecutorService, not Thread) is a static field in ZkBucketDataAccessor. The executor is started when ZkBucketDataAccessor class is initialized but it is never shut down. Since ExecutorService threads are generally not daemon threads, not shutting down GC_THREAD prevents JVM from shutting down cleanly.
This commit makes ZkBucketDataAccessor GC_THREAD a daemon thread so it doesn't prevent application shutdown.
Brent [Fri, 14 Jan 2022 19:57:57 +0000 (11:57 -0800)]
Upgrade Log4j to 2.16.0 to address CVE-2021-44228 (#1922)
* HELIX-1921: Upgrade Log4j to 2.16.0 to address CVE-2021-44228
- Upgrade SLF4J API version from 1.7.25 to 1.7.32 (latest)
- Remove use of slf4j-log4j12 package
- Add use of log4j-slf4j-impl package
- Remove unused custom log appender class
- Change direct Log4J reference to SLF4J
- Add -Dlog4j2.formatMsgNoLookups flag to scripts.
- Rename properties files to log4j2.properties and change CLI parameter to log4j2.configurationFile for Log4j2's precedence behavior
- Change properties files to use log4j2 syntax
- Add -Dlog4j2.configurationFile=file://"$BASEDIR"/conf/log4j2.properties to CLIs that were missing it
Qi (Quincy) Qu [Tue, 11 Jan 2022 19:48:54 +0000 (11:48 -0800)]
Improve helix tutorial and code formatting (#1931) (#1932)
Improve helix tutorial and code formatting
Jiajun Wang [Fri, 7 Jan 2022 22:21:45 +0000 (14:21 -0800)]
Avoid NPE when getting property store through Helix-rest API. (#1929)
This PR aims to fix the ambiguous error return message when user request to get an empty ZK node through the Helix-rest property store access API.
This PR changes the server behavior to response NO_CONTENT instead of internal_server_error in the scenarios described above.
Ramin Bashizade [Wed, 5 Jan 2022 19:45:54 +0000 (11:45 -0800)]
Declare dependency to Zookeeper in zookeeper-api-*.ivy (#1926)
Adds dependency to Zookeeper in the ivy file in zookeeper-api module.
xyuanlu [Mon, 20 Dec 2021 18:28:51 +0000 (10:28 -0800)]
Fix a string operation for custom health check and update test (#1924)
xyuanlu [Wed, 15 Dec 2021 21:41:36 +0000 (13:41 -0800)]
Add take/free instance implementation and test (#1918)
* take & free single instance impl
Jiajun Wang [Tue, 14 Dec 2021 00:51:39 +0000 (16:51 -0800)]
Make theadpool shutdown timeout configurable for the HelixTaskExecutor. (#1920)
Add TestHelixTaskExecutor.testHandlerResetTimeout() to cover the new changes.
Also refactoring the related code to reduce duplicate and confusing code.
xyuanlu [Fri, 3 Dec 2021 23:32:21 +0000 (15:32 -0800)]
Implement RestSnapShot and substitute the kv maps in HelixDataAccessorWrapper to a RestSnapShot object (#1913)
* implement RestSnapShot and substitute the kv maps in HelixDataAccessorWrapper with RestSnapShot object
xyuanlu [Thu, 2 Dec 2021 23:53:12 +0000 (15:53 -0800)]
Add rest API for take/free instance (#1917)
* add rest API for take/free instance
xyuanlu [Thu, 25 Nov 2021 03:05:04 +0000 (19:05 -0800)]
refactor instanceService to clusterMaintenanceService (#1912)
Junkai Xue [Mon, 22 Nov 2021 19:48:09 +0000 (11:48 -0800)]
Fix No Instance Level Throttling (#1908)
Instance level throttling quota never charged. Add the charging logic and tests.
xyuanlu [Mon, 22 Nov 2021 17:45:24 +0000 (09:45 -0800)]
Split BatchGetInstancesStoppableChecks (#1902)
Split BatchGetInstancesStoppableChecks into 2 private util functions.
Junkai Xue [Sun, 21 Nov 2021 21:15:56 +0000 (13:15 -0800)]
Add 0.9.9 to menu bar
Junkai Xue [Sun, 21 Nov 2021 20:27:19 +0000 (12:27 -0800)]
Missing end quote
Junkai Xue [Sun, 21 Nov 2021 20:09:42 +0000 (12:09 -0800)]
Fix site.xml head/footer
Junkai Xue [Sun, 21 Nov 2021 20:03:21 +0000 (12:03 -0800)]
Remove 0.9.7 doc folder
Junkai Xue [Sun, 21 Nov 2021 20:01:41 +0000 (12:01 -0800)]
Remove 0.9.1, 0.9.4, 0.9.7, 1.0.0 docs to keep only latest 2 releases doc
Junkai Xue [Sun, 21 Nov 2021 19:58:11 +0000 (11:58 -0800)]
Remove 0.8 series doc
Hervé Boutemy [Sun, 21 Nov 2021 10:18:01 +0000 (11:18 +0100)]
upgrade maven-site-plugin: fix site.xml head/footer (#1910)
Hervé Boutemy [Sat, 20 Nov 2021 11:42:23 +0000 (12:42 +0100)]
update parent and site plugin (#1909)
Junkai Xue [Wed, 17 Nov 2021 21:55:26 +0000 (13:55 -0800)]
Add additional note for release
xyuanlu [Wed, 17 Nov 2021 01:10:57 +0000 (17:10 -0800)]
fix input issue for stoppable rest API (#1905)
Junkai Xue [Tue, 16 Nov 2021 22:26:14 +0000 (14:26 -0800)]
Add 0.9.9 release notes
xyuanlu [Mon, 15 Nov 2021 19:41:23 +0000 (11:41 -0800)]
add take/free instance(s) API (#1899)
Create Cluster Maintenance Management service and add API signature and interfaces
Jiajun Wang [Thu, 4 Nov 2021 18:52:05 +0000 (11:52 -0700)]
Add additional ZK serializer configuration to active ZNRecord compression even the node size is smaller than write size limit.
The property zk.serializer.znrecord.auto-compress.threshold.bytes defines a threshold of ZNRecord size in bytes that the ZK serializer starts to auto compress the ZNRecord for write requests if it's size exceeds the threshold.
If the threshold is not configured or exceed ZKRecord write size limit, default value zk.serializer.znrecord.write.size.limit.bytes (if configured) or 1MB (if no configuration) will be applied.
xyuanlu [Mon, 1 Nov 2021 20:11:51 +0000 (13:11 -0700)]
Fix log format in instanceValidationUtil (#1894)
* fix log format in instanceValidationUtil
Junkai Xue [Mon, 27 Sep 2021 20:23:23 +0000 (13:23 -0700)]
Fix test failure TestInstancesAccessor (#1881)
Junkai Xue [Fri, 24 Sep 2021 17:46:23 +0000 (10:46 -0700)]
Fix inconsistent behavior beween batch stoppable and single stoppable API (#1879)
Neal Sun [Tue, 21 Sep 2021 23:15:55 +0000 (16:15 -0700)]
Fix TestClusterAggregateMetrics (#1842)
Fix TestClusterAggregateMetrics
Neal Sun [Tue, 21 Sep 2021 23:15:16 +0000 (16:15 -0700)]
Improve Purge Offline Instances API (#1870)
This commit improves the API such that it will also purge any incomplete instance data, such as instance path without InstanceConfig or ParticipantHistory.
Ali Reza Zamani Zadeh Najari [Fri, 17 Sep 2021 19:56:20 +0000 (12:56 -0700)]
Fix adding a task to a job after deleting old tasks (#1875)
In this commit, the issue of dynamically adding a task to a job
in which some of its tasks have been deleted before is being
addressed.
Huizhi Lu [Tue, 14 Sep 2021 18:15:55 +0000 (11:15 -0700)]
Fix flaky test testClusterFreezeMode (#1871)
The test fails randomly. Root cause is the test used a random cluster from the clusters set which has super clusters and task cluster, so the cluster could be a task cluster and then the test fails. Freeze mode does not apply to task framework.
This commit fixes it by changing the test cluster name with a fixed name TestClusters_0.
Junkai Xue [Tue, 14 Sep 2021 02:43:24 +0000 (19:43 -0700)]
Fix error partition blocks load rebalance (#1867)
There are three things fixed:
1. State priority is higher priority with smaller number.
2. When only downward is allowed, any non downward STs must be removed from message and throttled.
3. Even for downward STs should be respect to the throttling as backward compatible behavior.
4. Fix test for TestErrorReplicaPersist
Junkai Xue [Fri, 10 Sep 2021 00:06:07 +0000 (17:06 -0700)]
Change our release voting process
Acked from our Apache Helix VP, we will count committers' vote as binding vote for release.
xyuanlu [Wed, 8 Sep 2021 21:05:03 +0000 (14:05 -0700)]
Add input validation for getJobContext (#1864)
* add validation for getJobContext
Jiajun Wang [Wed, 1 Sep 2021 17:43:07 +0000 (10:43 -0700)]
Fix a potential race condition in MBean unregister logic.
The unregister method should relies on the MBeanServer class to validate if the target MBean has been unregistered or not to avoid race condition.
Jiajun Wang [Tue, 31 Aug 2021 21:01:10 +0000 (14:01 -0700)]
Upgrade mockito lib version to avoid test case failure due to mock mechanism legacy issues.
Jiajun Wang [Tue, 31 Aug 2021 17:24:05 +0000 (10:24 -0700)]
Improve TestControllerLeadershipChange test logic to tolerate longer delay when test runs. (#1853)
Change the assert condition to consider real test script execute delay to avoid test failure due to slow test runs.
Junkai Xue [Wed, 1 Sep 2021 19:07:53 +0000 (12:07 -0700)]
Remove unpublished 0.9.9
Junkai Xue [Wed, 1 Sep 2021 19:05:17 +0000 (12:05 -0700)]
Remove archived 0.6 & 0.7 version docs
Junkai Xue [Wed, 1 Sep 2021 18:36:04 +0000 (11:36 -0700)]
Update Menu Bar
Huizhi Lu [Wed, 1 Sep 2021 17:00:34 +0000 (10:00 -0700)]
Fix management mode history duplicate recording (#1846)
The management mode history has duplicate entries. It does not impact the normal function, but it's good to get it fixed to avoid confusion. This commit fixes the issue by adding a check for the status in metadata store and the calculated status.
Jiajun Wang [Tue, 31 Aug 2021 20:02:33 +0000 (13:02 -0700)]
Revert "Improve TestControllerLeadershipChange test logic to tolerate longer delay when test runs. (#1853)" (#1860)
This reverts commit
a570d0566c42942b6154cb84a3d44f864fde37f0.
Jiajun Wang [Tue, 31 Aug 2021 17:24:05 +0000 (10:24 -0700)]
Improve TestControllerLeadershipChange test logic to tolerate longer delay when test runs. (#1853)
Change the assert condition to consider real test script execute delay to avoid test failure due to slow test runs.
xyuanlu [Mon, 30 Aug 2021 18:25:34 +0000 (11:25 -0700)]
Throw exception when query partition assignment in Maintenance mode (#1855)
Throw exception for partitionAssignment when cluster in Maintenance mode.
xyuanlu [Mon, 30 Aug 2021 18:25:06 +0000 (11:25 -0700)]
Disable delayed rebalance as default for partitionAssignment (#1852)
Disable delayed rebalance as default for partitionAssignment.
xyuanlu [Tue, 24 Aug 2021 21:23:45 +0000 (14:23 -0700)]
Use activate/DeactivateInstances as keyword in PartitionAssignment API (#1850)
Use activate/DeactivateInstances as keyword in PartitionAssignment API.
Jiajun Wang [Tue, 24 Aug 2021 00:42:20 +0000 (17:42 -0700)]
Add stop server wait time to be 10 seconds in MockMetadataStoreDirectoryServer to avoid test failure. (#1848)
This change aims to reduce the unexpected test failure due to http endpoint in use. The additional timeout shall help the test terminate server instances gracefully.
Jiajun Wang [Mon, 23 Aug 2021 23:04:20 +0000 (16:04 -0700)]
Check server field state before shutdown to avoid NPE in MockMetadataStoreDirectoryServer. (#1847)
Check server field state before shutdown to avoid NPE in MockMetadataStoreDirectoryServer.
Junkai Xue [Thu, 19 Aug 2021 19:55:36 +0000 (12:55 -0700)]
Remove duplicated notice content
Junkai Xue [Thu, 19 Aug 2021 19:53:38 +0000 (12:53 -0700)]
Update Lisence and Notice
Neal Sun [Mon, 16 Aug 2021 18:22:49 +0000 (11:22 -0700)]
Add TF Available Threads Metrics (#1834)
Add metrics about Task Framework available threads in the cluster per job type.
Junkai Xue [Mon, 9 Aug 2021 18:44:07 +0000 (11:44 -0700)]
Fix test logic for testExternalViewDiffFromTargetExternalView (#1835)
Ali Reza Zamani Zadeh Najari [Thu, 5 Aug 2021 18:35:01 +0000 (11:35 -0700)]
Stabilize TestInstancesAccessor (#1828)
TestInstancesAccessor is unstable because the cluster might not be
stable while the test runs the stoppable check and cluster config and
instance config can be missing. In this commit, the test waits until the
cluster config and instance config get created and then proceeds for
the stoppable checks.
Jiajun Wang [Wed, 4 Aug 2021 19:33:45 +0000 (12:33 -0700)]
Fix the JavaDoc about transition priority. (#1831)
Fix the JavaDoc about transition priority.
In addition, update several old code style for simplicity.
Neal Sun [Tue, 3 Aug 2021 17:57:12 +0000 (10:57 -0700)]
Fix TestZkConnectionLost (#1824)
Fix TestZkConnectionLost by adding longer connection timeout and fixing logic that are incorrect.
Co-authored-by: Neal Sun <nesun@nesun-mn1.linkedin.biz>
xyuanlu [Tue, 3 Aug 2021 17:53:59 +0000 (10:53 -0700)]
Add response metadata to response header for partitionAssignment (#1797)
Add response metadata to response header for partitionAssignment.