Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

Table of Contents

 

Status

Current stateUnder discussionAccepted.

Discussion threadhere

Vote thread: here

JIRAKAFKA-3487

Released:  0.11.0.0

Please keep the discussion on the mailing list rather than commenting on the wiki (wiki discussions get unwieldy fast).

Motivation

Kafka Connect is a highly extensible framework, designed to run software modules that integrate a wide variety of data sources and destinations with Kafka and, by definition, are developed independently of the framework. Examples of such software modules are connectorstransformations and data converters, which hereafter will be called plugins to honor Kafka Connect terminology. Because such independent development, dependency conflicts might arise naturally, between the framework and the modulesplugins, as well as among the modules plugin themselves. This happens when, for instance, one connector is using the latest version of a package that added a new feature, while another connector depends on a previous version of the same package where this feature is missing. 

...

Nevertheless, as Kafka Connect adoption grows, runtime exceptions due to version mismatches might arise more often, and therefore present could become a burden in the development and deployment of connectors with Kafka Connect. 

...

In this case, the framework will do the heavy lifting in terms of development, offering a transparent resolution of dependency conflicts and keeping the changes in Kafka Connect's public interfaces minimal. 

Public Interfaces

The only publicly visible changes change that are is proposed in order to implement class loading isolation in Kafka Connect is the addition of two a new Connect worker config propertiesproperty:

This is a These are framework level configuration properties property that will affect all the connectors running in within a Connect worker. 

The new configuration option moduleproperty plugin.path will accept a list of locations, represented as strings and separated by commas. The strings representing locations should be able to be transformed into URLs , in order to offer maximum flexibility in terms of module plugin discovery. Nevertheless, in the first implementation of class loading isolation such locations will be paths to the local filesystem, that, in turn, can be easily transformed into URLs. 

Examples of valid values for plugin.path:

  • plugin.path=/usr/local/share/java
  • plugin.path=/usr/share/java, /usr/local/share/java, /opt/connectors

When a filesystem path is used as location for imported modules plugins (connectors, transformations and converters), the convention is that each module plugin is storing all its required dependencies in a single directory under the each location path listed in moduleplugin.pathFor instance, if the location /usr/local/share/java is given in moduleplugin.path then modules plugins such as my-kafka-source-connectormy-kafka-sink-connectormy-connect-smt and my-converter should store all of their dependencies (usually in the form of jar files, but possibly as raw class files with the appropriate java package structure) under directories as immediately below each plugin.path entry as follows:

/usr/local/share/java

/my-kafka-source-connector/(connector jars and dependency jars)

/my-kafka-sink-connector/(connector jars and dependency jars)

/my-connect-smt/(smt jars and dependency jars)

/my-converter/

The new configuration option module.isolation.enabled will accept a boolean value that will turn the feature of class loading isolation on and off. 

(converter jars and dependency jars)

A more nested storing of plugin and dependency jars will not make them discoverable.

Again with plugin.path set to /usr/local/share/java an incorrect example would be: 

 /usr/local/share/java

/more-plugins/another-kafka-source-connector/(connector jars and dependency jars)

In this case, in order to make this plugin discoverable,  /usr/local/share/java/more-plugins should be added to plugin.path instead.

Additionally, specifically regarding Connect plugins, each connector, transformation or converter jar should be listed only once across the plugin path. If multiple copies of the same version of plugins exist under plugin.path the selection of which plugin will be loaded will be deterministic but implementation specific.

While the introduction of the new configuration property plugin.path is the only change While the introduction of these two new configuration properties are the only changes proposed to the public interfaces of Kafka Connect, next are also described the main implementation steps that will be carried out within the framework to support class loading isolation.

Proposed Changes

Add moduleplugin.path and module.isolation.enabled configuration properties property for workers running both in standalone and distributed mode. By default moduleplugin.path is empty and module. In this case isolation .enabled is set to false.is not active and loading of connectors depends on the CLASSPATH. In general, when a user wants to run a connector is isolation, its packages along with its dependencies should be stored under a directory that is listed in plugin.pathOtherwise if non-isolated execution is desired, the connector jars should be listed in the CLASSPATH. 

Given that plugin.path is set appropriatelyBased on those two configuration properties, when class loading isolation is enabled, the Connect framework will be able to instantiate a custom module plugin classloader for each module plugin under the list of locations supplied in the moduleplugin.pathThe main characteristics of such a module plugin classloader are that:

  • it filters out classes belonging to the java library and the Connect framework and delegates their loading delegates loading of Connect framework classes as well as java library classes to the appropriate classloader. Optionally, warnings could be issued when a plugin path contains Connect framework classes.
  • it applies a child-first policy for the rest of the classes, aiming to load module plugin specific dependencies directly.

...

  • the Connect framework controls the threads that run module plugin code (e.g. connector tasks).
  • module plugin classes and dependencies are required to be supplied explicitly through moduleplugin.path.

Compatibility, Deprecation, and Migration Plan

  • Existing users will not be impacted since isolation will be disabled by defaultan non-existent/empty plugin.path means that isolation is not in effect
  • Users enabling class loading isolation might experience higher demands in memory usage due to additional loading of otherwise common classes. However this increase is not expected to be prohibitive in most cases.

Test Plan

  • Targeted Unit tests will be developed to test the components that will implement class isolation in the framework. Additionally, all other tests will be set to run with class loading isolation turned on. 
  • System tests will be written to test class loading isolation when explicitly conflicting dependencies are introduced by connectors. 
  • Microbenchmarks will be designed to make sure the effect of classloader context switching is negligible

Future Work

In a subsequent version of class loading isolation, the following enhancements will be targeted: 

  • Support versioned plugins. Users will have the ability to load and run different versions of the same plugin. Although connectors are versioned already, this capability needs to be enabled for transformations and converters as well. 
  • Extend plugin.path to support plugin locations beyond the local filesystem. Such locations might include web addresses, packages repositories (e.g. maven repositories) and other locations

Rejected Alternatives

  • Use OSGi. OSGi. Along with its much broader scope, an OSGi implementation would bring significant implementation complexity, both in the framework and the connector development. 

  • Design for project Jigsaw and wait. With this KIP we describe an implementation path whose execution is expected to be focused and efficient. Major upgrades, such as an upgrade to the next Java version are not expected to move faster than the proposed implementation.

...