[KAFKA-13306] Null connector config value passes validation, but fails creation

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 3.2.0
• Component/s: connect
• Assignee: Laszlo Istvan Hunyady
• Reporter: Laszlo Istvan Hunyady
• Created: 16/Sep/21 12:28
• Updated: 11/Feb/22 15:14
• Resolved: 11/Feb/22 15:14

Description

When validating a connector config containing a property with a null value the validation passes, but when creating a connector with the same config the worker fails to start the connector because of an invalid config.

Steps to reproduce:

1. Send PUT request to
   connectRest/connector-plugins/FileStreamSource/config/validate
   Request body:

   {
     "connector.class": "FileStreamSource",
     "name": "file-source",
     "topic": "target-topic",
     "file":"/source.txt",
     "tasks.max": "1",
     "foo": null
   }
   

   Response:

   { "name": "FileStreamSource", "error_count": 0, ... }
   

1.  OPTION A:
   Send PUT request to
   connectRest/connectors/file-source/config
   Request body:

   {
     "connector.class": "FileStreamSource",
     "name": "file-source",
     "topic": "target-topic",
     "file":"/source.txt",
     "tasks.max": "1",
     "foo": null
   }
   

   OPTION B:
   Send POST request to
   connectRest/connectors/
   Request Body:

   {
     "connector.class": "FileStreamSource",
     "name": "file-source",
     "topic": "target-topic",
     "file":"/source.txt",
     "tasks.max": "1",
     "foo": null
   }
   

   Result:
   Connector is created but connector fails to start, with below exception that indicates an invalid config:

   ERROR org.apache.kafka.connect.runtime.WorkerConnector: [file-source|worker] WorkerConnector\{id=file-source} Error initializing connector
   java.lang.ClassCastException: Non-string value found in original settings for key foo: null
           at org.apache.kafka.common.config.AbstractConfig.originalsStrings(AbstractConfig.java:234)
           at org.apache.kafka.connect.runtime.WorkerConnector.initialize(WorkerConnector.java:77)
           at org.apache.kafka.connect.runtime.Worker.startConnector(Worker.java:256)
           at org.apache.kafka.connect.runtime.distributed.DistributedHerder.startConnector(DistributedHerder.java:1190)
           at org.apache.kafka.connect.runtime.distributed.DistributedHerder.processConnectorConfigUpdates(DistributedHerder.java:548)
           at org.apache.kafka.connect.runtime.distributed.DistributedHerder.tick(DistributedHerder.java:395)
           at org.apache.kafka.connect.runtime.distributed.DistributedHerder.run(DistributedHerder.java:289)
           at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
           at java.util.concurrent.FutureTask.run(FutureTask.java:266)
           at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
           at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
           at java.lang.Thread.run(Thread.java:748)
   

    

Issue Links

• links to: GitHub Pull Request #11333