Trying to port my Spark 2.4 based (Structured) streaming application to Spark 3.0. I compiled it using the dependency given below:

<dependency>
<groupId>org.apache.spark</groupId>
<artifactId>spark-sql-kafka-0-10_${scala.binary.version}</artifactId>
<version>3.1.0</version>
</dependency>

Every time I run it under Spark 3.0, I get this message: Data source v2 streaming sinks does not support Update mode

I am using 'mapGroupsWithState' so as per this link (https://spark.apache.org/docs/latest/structured-streaming-programming-guide.html#output-modes), the only supported Output mode is "Update". 

My Sink is a Kafka topic so I am using this:

.writeStream
.format("kafka")

What am I missing?